Task 2

Drainage Water Treatment

Final Report

February 1999

Drainage Water Treatment Technical Committee

The San Joaquin Valley Drainage Implementation Program

and

The University of California Salinity/Drainage Program

Please note that this technical reportis being considered by a publishing company

for possible publication in a book in the near future.

DISCLAIMER

This report presents the results of a study conducted by an independentTechnical Committee for the Federal-State Interagency San Joaquin Valley DrainageImplementation Program. The Technical Committee was formed by the University ofCalifornia Salinity/Drainage Program. The purpose of the report is to provide theDrainage Program agencies with information for consideration in updating alternativesfor agricultural drainage water management. Publication of any findings orrecommendations in this report should not be construed as representing theconcurrence of the Program agencies. Also, mention of trade names or commercialproducts does not constitute agency endorsement or recommendation.

The San Joaquin Valley Drainage Implementation Program was established in1991 as a cooperative effort of the United States Bureau of Reclamation, United StatesFish and Wildlife Service, United States Geological Survey, United States Departmentof Agriculture-Natural Resources Conservation Service, California Water ResourcesControl Board, California Department of Fish and Game, California Department of Foodand Agriculture, and the California Department of Water Resources.

For More Information Contact:

Manucher Alemi, CoordinatorThe San Joaquin Valley Drainage Implementation ProgramDepartment of Water Resources1020 Ninth Street, Third FloorSacramento, California 95814(916) 327-1630malemi@water.ca.govor visit the SJVDIP Internet Site at:http://wwwdpla.water.ca.gov/agriculture/drainage/implementation/hq/title.htm

Drainage Treatment Technical Committee Report

Outline

I. Regulatory Perspective of Treatment Technologies for SubsurfaceAgricultural Drainage Water………………………………………………….…..…..1

II. Treatment Technologies………………………………………………………..…….3

A. Ion Exchange……………………………………………………………….….3B. Distillation Processes……………………………………………….…….…..4C. Membrane Processes……………………………………………..……….…4D. Solar Ponds……………………………………………………………..……..9E. Chemical Reduction of Selenium with Zero-valent Iron………………….10F. Chemical Reduction of Selenium with Ferrous Hydroxides……………..11G. Algal-Bacterial Selenium Removal Process……………………………….11H. In Situ Treatment Utilizing Aquatic Algal Productivity………………….…14I. Algal Bioremediation Studies………………………………………………..15J. Selenium Removal from Water by Algal Processes……………………...16K. Other Current Work that Includes Algal Processes……………………....17L. Criteria for Success of In Situ Biologically Based Selenium

Treatment Technology……………………………………………………..17M. Volatilization…………………………………………………………………...18N. Biological Precipitation……………………...………………………………..20O. Flow-Through Wetlands……………………………………………………...22P. Subsurface Flow Wetlands………………………………………………..…24

III. Conclusions……………………………………………………………………..……..26

References………………………………………………………………………………..……28

Appendix………………………………………………………………………………………..36

San Joaquin Valley Drainage Implementation Program

Report from the Technical Committee on Drainage Water Treatment

Active Committee Members:

William Frankenberger, Jr. University of California, Riverside

Christopher Amrhein University of California, Riverside

Teresa Fan University of California, Davis

Dennis Falaschi Panoche Water District, Firebaugh, California

Julius Glater University of California, Los Angeles

Ernest Kartinen, Jr. Boyle Engineering Corporation

Kurt Kovac Department of Water Resources

Edwin Lee Private Consultant

Harry Ohlendorf CH2M Hill, Inc.

Larry Owens California State University, Fresno

Norman Terry University of California, Berkeley

Anthony Toto Central Valley Regional Water Quality Control Board

Advanced Treatment Technologies in the Remediation of SeleniferousDrainage Waters and Sediments

W. T. Frankenberger, Jr., C. Amrhein, T. W. M. Fan, D. Flaschi, J. Glater,E. Kartinen, Jr., K. Kovac, E. Lee, H. M. Ohlendorf,

L. P. Owens, N. Terry, and A. Toto

This review identifies and evaluates research findings on various treatmentoptions for selenium removal and disposal of salts in the San Joaquin Valley, California.Selenium from agricultural drainage water has the potential to bioaccumulate at highenough levels in plants and animals so as to cause mortality and to impair reproductionof fish and aquatic birds. The selenium link between agricultural drainage and impactson birds was first documented in 1983 by U.S. Fish and Wildlife Service studies atKesterson Reservoir in the SJV. Aquatic plants, invertebrates, fish, frogs, snakes, birdsand mammals at Kesterson Reservoir contained elevated selenium levels, oftenaveraging a 100-fold increase over samples collected for similar species at referencesites (Ohlendorf, 1989; Ohlendorf and Santolo, 1994).

I. REGULATORY PERSPECTIVE OF TREATMENT TECHNOLOGIES FORSUBSURFACE AGRICULTURAL DRAINAGE WATER

The treatment of subsurface agricultural drainage water has several regulatoryconsiderations. The goals or objectives of treatment technologies can be categorizedas: (i) reduce constituents below hazardous levels, (ii) meet agricultural water qualitygoals, (iii) meet water quality objectives in surface waters, and (iv) reduce constituentsbelow risk levels for wildlife.

The first goal of drainage water treatment is to reduce constituents belowhazardous levels. The two most common elements in drainage water that approachhazardous levels are selenium and arsenic. In some subareas (e.g., Westlands),drainage water contains levels of selenium that approach the numeric criterion defininghazardous waste. This level is defined in Title 22, California Code of Regulations.Selenium levels above 1,000 µg/L and arsenic levels above 5,000 µg/L are consideredhazardous. The use of treatment technologies to reduce constituent levels belowhazardous criteria would be appropriate in: (i) areas with high selenium levels;(ii) areas with high arsenic levels; (iii) areas where reuse of drainage water(e.g., agroforestry) concentrates trace elements above numeric criteria defininghazardous waste; and (iv) where required to allow disposal of the water to a river orstream.

The second goal for meeting agricultural water quality standards is applicable toall subareas. Treatment of drainage water to be reused for agricultural supply wouldrequire reducing boron (700 µg/L) and salt (TDS = 450 mg/L and EC = 700 µS/m) levelsto meet agricultural water quality goals (Ayers and Westcot, 1985) or reduce the boronand salt levels sufficiently to allow blending with a better quality water supply. Theseagricultural water quality goals would not be applicable as a regulatory number tomanage drainage at the farm level.

2

The third goal of meeting water quality objectives is applicable to surface waterdischarges. The primary discharge of drainage water is to the San Joaquin River,California, although some drainage water is discharged to the Kings River and othersurface waters. Water quality objectives for selenium and salts are established for theSan Joaquin River and listed in The Water Quality Control Plan for the SacramentoRiver Basin and the San Joaquin River Basin, Third Edition and Amendments (RWQCB,1994). In the year 2010, the water quality objective in the San Joaquin River forselenium concentration will be 5 µg/L (based on 4-day average). The objective forsalinity in the San Joaquin River at Vernalis is an electrical conductance (based upon a30-day running average) of 700 µS/m from April to August and 1,000 µS/m Septemberto March. Also, drainage discharge to the Kings River in the Tulare Lake Basin mayoccur provided water quality objectives are met and beneficial uses are not impaired(The Water Quality Control Plan for the Tulare Lake Basin, 1995). Discharge to theKings River must have an electrical conductance below 1,000 µS/m, boron level below1 mg/L, and chloride level below 175 mg/L.

The fourth goal, reducing constituents below risk levels for wildlife, would apply toprimarily any open body of water containing elevated levels of selenium (>2 µg/L). Thisstandard would include not only evaporation basins but also drainage water storagebasins. The subareas would include: (i) the Tulare and Kern subareas for evaporationbasins and (ii) the Grasslands subarea for drainage water storage basins. Treatmenttechnologies would need to reduce the selenium levels to about 2 µg/L or isolate themechanism for bioaccumulation to eliminate risk to wildlife associated with selenium.Currently, operators of evaporation basins with a geometric mean seleniumconcentration greater then 2 µg/L are required to analyze egg selenium levels (eggselenium is a better indicator of wildlife impact than water selenium). If egg seleniumlevels are elevated then mitigation habitat must be provided. Selenium treatmenttechnologies must consider the potential of mitigation habitat when selenium levelsexceed 2 µg/L.

Treatment technologies currently being evaluated to reduce the selenium load indrainage water include physical, chemical and biological methods. There are threerecent books that provide worldwide compilation of scientific studies related to thetreatment technologies of selenium, edited by Jacobs et al. (1989), Frankenberger andBensen (1994), and Frankenberger and Engberg (1998). A detailed discussion ofdrainwater treatment methods, field trials and relative costs was the focus of a report byHanna et al. (1990), submitted to the San Joaquin Valley Drainage Program,Sacramento, California. This report will focus on ion exchange, reverse osmosis, solarponds, chemical reduction with iron, microalgal-bacterial treatment, volatilization,biological precipitation, and flow-through wetlands as viable technologies.

3

II. TREATMENT TECHNOLOGIES

A. Ion Exchange

Ion exchange involves the exchange of an undesirable dissolved constituent for amore desirable solute electrostatically attached to an ion exchange material. Thismaterial is often a synthetic resin, but for the purposes of this discussion may alsoinclude a naturally occurring zeolite or activated alumina. While ion exchange is usedfor desalting water (deionized water is commonly used in industry), it is generallyuneconomic for treatment of agricultural drainage when compared with reverseosmosis. In fact, most industries have upgraded their deionized water facilities byproviding reverse osmosis as a pretreatment, followed by an ion exchange system topolish the water to extremely low TDS levels. However, ion exchange is often used forspecific ion removal rather than overall salinity reduction. Examples include the nitrateremoval facilities found at McFarland and other California cities. Ion exchange systemsfind widespread use for water softening in the removal of calcium and magnesium(hardness reduction).

Ion exchange resins can selectively remove a specific ion in preference to others.Recent research has shown that resins are suitable for targeting arsenic, selenium andother ionic constituents. However, since selenate (the common form of selenium indrainage water) and sulfate have very similar chemical properties, it is difficult toseparate selenate (at the parts per billion concentration) from sulfate (at the parts permillion concentration) since a resin would quickly become saturated with sulfate ionsand stop removing selenate. Research performed at the Panoche Water District hasdemonstrated that at least one commercially available resin is capable of removingselenate ion even when saturated with sulfate. The selenate ion was primarily removedat low concentrations.

Sometimes, selective ion removal can be accomplished by properly pretreatingthe drainwater. For example, arsenic can be present in several ionic forms, dependingupon the pH and oxidation state (pΕ) of the water. At certain pH and pΕ ranges, arsenicis in the arsenious acid form, and will not be removed by ion exchange. When the wateris oxidized (e.g. by addition of chlorine) and the pH properly adjusted, arsenic isconverted to arsenate ion, which is readily removed by ion exchange in preference tomany other common ions. Thus, appropriate pretreatment may be the key toeconomical application of using ion exchange for selective removal of a specificcontaminant.

Ion exchange is a very simple process to operate and is suitable for intermittentflow. Labor requirements are low, typically requiring an hour a day to monitor thecondition of the equipment. Ion exchange produces a salty waste stream, typically to1 to 5 percent of the product water. This waste contains material removed from thetreated water, as well as unused salt from the regeneration process. Disposal oftenrequires evaporation and landfilling of the precipitated salts, although it is possible thatin certain cases the waste may be useable for some beneficial purpose. When used forspecific ion removal, ion exchange is relatively inexpensive on both a capital andoperating cost basis when compared with other processes. The regenerable nature ofthe treatment medium helps to distribute the purchase cost over the long term. An ion

4

exchange treatment system can be expected to have a capital cost in the range of$0.30 to $0.50 per gallon per day capacity, depending upon the need for specializedpretreatment. Operating and maintenance costs can be expected to range from $30 to$100 per acre foot of water.

The capital cost of ion exchange equipment is less than for reverse osmosis ordistillation equipment. However, the cost of regenerating chemicals for waters with aTDS over a few hundred mg/L generally make ion exchange economically unattractivefor demineralizing water when the “total water cost” (capital and operation andmaintenance costs) are considered. But, the use of ion exchange for specific ionremoval, such as selenium, holds promise and should be considered.

B. Distillation Processes

Distillation processes, as the name implies, rely on evaporation andcondensation to demineralize water. While distillation processes are used to desalt seawater, energy consumption is substantially more, for example, than sea water desaltingusing reverse osmosis. In addition, the capital cost of distillation plants is usually morethan for reverse osmosis. The cost of distillation would be prohibitive when applied tobrackish irrigation drainage waters since energy consumption is not related to salinity.For example, distillation of seawater at 35,000 mg/L would cost essentially the same asapplication of this treatment to drainage water at less than 10,000 mg/L.

Distillation processes are mentioned here because they could be used todemineralize irrigation drainage water. However, as noted above, they are almostwithout exception more expensive than reverse osmosis except for locations whereenergy is inexpensive, such as in Saudia Arabia, where large supplies of natural gasare available and where sea water desalting plants have been built as part of a “dualpurpose” facility; i.e., power generation and water production.

C. Membrane Processes

There are several membrane water treatment processes:

• Microfiltration—a solids (particulate) removal filtration process with membrane poresizes of about 0.2 µm. Microfiltration does not remove dissolved substances fromwater but can be used to pretreat (remove solids) from water for further treatment bya desalting process.

• Ultrafiltration—a solids removal filtration process similar to microfiltration except themembrane pore size is about 0.02 µm.

• Nanofiltration—a reverse osmosis process sometimes called “membrane softening.”Nanofiltration will remove dissolved substances from water and is especiallyeffective for removing divalent ions such as calcium, magnesium and sulfate hencethe name “membrane softening”.

5

• Reverse osmosis (hyperfiltration)—a dissolved substances removal process that isalso very effective at removing monovalent ions, such as sodium, chloride, etc., aswell as divalent ions from water.

• Electrodialysis—a membrane process that is used to remove electrically chargeddissolved substances from water.

Micro- and ultrafiltration are solids removal membrane processes and thisdiscussion concerns removal of dissolved substances from water. They are mentionedhere only because they might find application in pretreating irrigation water by adissolved solids removal process (reverse osmosis, ion exchange, or nanofiltration).

There are two primary differences between the reverse osmosis processes(nanofiltration and hyperfiltration) and electrodialysis; what passes through themembranes and the driving force. In reverse osmosis the water passes through themembranes. The “rejected” salts (salts and other dissolved substances) remain on thefeed water side of the membrane (the “concentrate side”). The water passes throughthe membrane to the “permeate side” The driving force is pressure.

In electrodialysis, the electrically charged ions pass through the membranes, notthe water as in reverse osmosis. The driving force in this case is electric potential(voltage).

While both reverse osmosis and electrodialysis can be used to demineralizeirrigation drainage water, electrodialysis is usually somewhat more expensive thanreverse osmosis, especially when the TDS exceeds about two or three thousand mg/L.Energy consumption of both processes is directly related to the feed water salt content.Studies have shown, given environmental constraints and the current level oftechnology, that reverse osmosis usually provides lower costs for desalting brackishwater. In previous demonstrations with agricultural drainage, reverse osmosis has beenfavored over electrodialysis.

Reverse osmosis is applicable to natural waters over a wide salinity rangewhereas electrodialysis is generally restricted to salinity levels of 15,000 mg/L or less.Energy consumption of both processes is directly related to the feed water salt content.Studies have shown, given environmental constraints and the current level oftechnology, that reverse osmosis provides substantially lower costs for desalting evenseawater. For brackish water, this process is especially attractive. In previousdemonstrations with agricultural drainage, reverse osmosis has been favored overelectrodialysis.

Reverse osmosis is a "broad spectrum" treatment process capable of removingnumerous contaminants. In this technology a semipermeable membrane selectivelyallows the passage of solvent (water) while rejecting the solute (dissolved salts andorganics). Osmosis is a term describing the natural tendency of a system to equalizesolute concentration on both sides of a semipermeable membrane as solvent migratesthrough this barrier from low to higher solute concentration. Reverse osmosis, then,reverses the natural osmotic process by applying pressure to the solution with highersolute concentration. In this manner, pure solvent is forced through the membrane.

6

Pressures required to drive this process must exceed the osmotic pressure of the salinewater being treated.

Early reverse osmosis membranes were developed from asymmetric celluloseacetate films (Loeb and Sourirajan, 1960). This material is still in use, however,performance is limited and cellulosic derivatives are easily fouled by bacterial growth onthe membrane surface. Other polymeric materials and fabrication techniques rapidlyfollowed the “first” reverse osmosis membrane (Petersen, 1993). These “newgeneration” membranes show dramatically higher levels of salt rejection and productwater recovery at a given driving pressure. A widely used class of high performancemembranes are generally referred to as thin film composite (TFC). They are formed bycasting a thin polymeric film onto a porous substrate composed of several layers ofvarious materials. Many different types of TFC membranes are available and are ratedaccording to their pressure requirements, flux characteristics and capability of saltrejection (typically 95 to 99.7 percent).

Higher rejection membranes generally require higher pressures to provide agiven product water flux. However, great strides have been made during the past fewyears in improving the pressure-flux characteristics of membranes while maintainingexcellent salt rejection. A recently developed high performance TFC type of membraneknow as nanofiltration (NF) has shown some interesting performance characteristics.These films, also known as softening membranes, operate at remarkably low pressures,providing reasonable overall salt rejection but have been designed to selectively rejectdivalent ions.

While there are differences in rejection of various ions, they are relatively minor.For example, a given membrane may show a rejection of 99 percent for nitrate but only96 percent for chloride. Membranes generally reject multivalent ions better thanunivalent and any charged particle better than an uncharged species. Rejection ofdissolved organic compounds can vary widely and depends on the size and structure ofthe organic molecule.

If the dissolved ions in the water are primarily divalent ions (calcium, magnesium,and sulfate) nanofiltration membranes may be better suited for desalination ofagricultural drainage waters than reverse osmosis membranes. If, however, there is alarge percentage of monovalvent ions in the water (sodium and chloride, for example)then reverse osmosis membranes may be more suitable. While nanofiltrationmembranes used to have a substantial advantage over reverse osmosis membraneswith respect to operating pressure (energy consumption), the development of “ultra lowpressure” reverse membranes in the last few years has largely narrowed the formerenergy consumption advantage of nanofiltration membranes.

Depending on the membrane type, salt rejections up to 99 percent are routinelyobserved with reverse osmosis. If the TDS of the irrigation water is, for example,10,000 mg/L, the permeate may have a TDS of only about 100 mg/L. Membranesgenerally reject multivalent ions better than univalent and any charged particle betterthan an uncharged species. Rejection of dissolved organic compounds can vary widelyand depends on the size and structure of the organic molecule.

7

The impact of membrane fouling must also be carefully considered in the designof any membrane desalination plant. Fouling may result from feed waters containingparticulate matter, potential scale forming ions, dissolved organics or micro-organisms.Depending on the source water quality, design of pretreatment systems is of crucialimportance.

The recovery (percent of feed water “recovered” as permeate) from a membraneplant depends on several factors including water chemistry, membrane configuration,and design of pretreatment systems. The latter factor is of crucial importance sincemost agricultural drainage waters contain very high levels of calcium and sulfate ionswhich can readily precipitate gypsum (CaSO4Α2H20) on membrane surfaces as productwater recovery increases. Addition of scale inhibiting chemicals to feed water providesfor operation at higher rates of recovery.

A reverse osmosis plant treating drainage water with TDS levels below about10,000 mg/L can be expected to have a capital cost of about $2.00 per gallon per day ofcapacity. If extensive pretreatment (to remove solids from the reverse osmotic feedwater) is required, the cost may rise an additional $1.00 per gallon per day of capacity.Operating costs can be expected to range between $150 and $300 per acre footdepending on feed water quality. A recent study on drainage water desalting costscompleted for the Buena Vista Water Storage District developed a water cost of about$300 per acre foot, including capital repayment (20 years at 7 percent) and O&M cost.This figure did not include the cost of facilities to gather the drainage water, dispose ofthe waste, or deliver the treated water. In another study on softening of well water inFlorida (Bergman, 1996), costs of nanofiltration membrane treatment ranged between$2.05 and $3.05 per 1,000 gallons. This overall cost included amortized constructionand O&M.

The first attempt at drainage water reclamation began in 1971 at Firebaugh,California (McCutchan et al., 1976). A small membrane pilot plant utilizing hand-castcellulose acetate tubular membranes was designed and built at theUCLA School of Engineering and Applied Science. The plant remained on-line forapproximately three years and was operated jointly by UCLA and the CaliforniaDepartment of Water Resources. Water quality at this site varied in TDS levelsbetween 2,000 and 7,000 mg/L and calcium and sulfate ion concentrations nearsaturation with respect to gypsum. A limiting issue in processing this water was thepotential deposition of scale on membrane surfaces. Scale control was effected first bytreatment with sodium hexametaphosphate followed by installation of a cation exchangesystem for calcium removal. Product water recovery based on chemical and ionexchange treatment, were reported at 60 percent and 90 percent respectively.Adequate levels of product water flux and overall salt rejection were also reported butusual operating pressures ranged between 600 and 900 psi.

Feasibility of this technology in drainage water reclamation was clearlydemonstrated at Firebaugh but cost effectiveness was limited by the very high pressurerequirements of the cellulose acetate membranes available at that time. It should benoted that the advancement in reverse osmosis membrane technology in the last fewyears has resulted in membranes that operate at pressures of about 1/3 of thoseexperienced at Firebaugh. It should also be noted that construction and operation of ion

8

exchange systems would place another large economic constraint on membranedesalination of SJV drainage water.

The next phase in membrane applications took place at Los Banos between1983 and 1987 (Smith, 1992). This sophisticated installation incorporated a number ofnovel pretreatment strategies and was tied in with solar pond development whichprovided a “sink” for brine disposal and utilization of solar energy. Membranes installedat this facility were provided by three leading manufacturers and represented typicalreverse osmosis (RO) technology available at that time. Regretfully this site was shutdown in 1986 following the discovery of environmentally hazardous levels of selenium atthe Kesterson Reservoir. Although the reverse osmosis units operated for relativelyshort periods, overall performance, once again, demonstrated the potential feasibility ofmembrane desalination for agricultural drainage water treatment.

Further studies on membrane desalination with “new generation” nanofiltration(NF) membranes show high levels of salt rejection while operating at remarkably lowpressures. Performance parameters of these membranes have been evaluated with“model solutions” simulating the composition of drainage water at the Adams AvenueAgricultural Drainage Water Research Center. At applied pressures less than 200 psi,product water flux (gallons per day per square foot of membrane area – GFD) and saltrejection were measured at levels up to approximately 18 GFD and 90 percentrespectively. Such data tend to forecast much lower product water costs compared withtechnology available in the early l980's.

Currently, research efforts are focused on potential gypsum scaling which is acritical issue in the design of pretreatment for a membrane desalination facility(LeGouellec et al., 1997). Work to date shows encouraging results with some of theleading commercial antiscalants. Goals are directed toward development of a pilot plantwith performance levels equal to or exceeding those measured in the laboratory andcapable of product water recovery at approximately70 percent.

At this time hard data is not available to adequately provide for an economicassessment of agricultural drainage water desalination by membrane processes butfollowing is a list of factors both positive and negative which would influence theplanning of membrane treatment facilities.

POSITIVE

1. NF membranes demonstrate remarkable efficiency for potential drainage watertreatment in comparison with the best membranes available in the early 1980'swhen design specifications for the Los Banos facility were under development.

2. Brine generated from membrane treatment will have an overall ionic concentrationapproximately three times ambient drainage water. The enhanced selenate levelsshould facilitate the efficiency of proposed bioremediation and phytoremediationschemes. However, the impact of increased salinity on biological processes ispresently unknown.

9

3. Reduced brine volumes will simplify the design of evaporating ponds and/or provideconcentrated solutions for solar ponds if this technology is re-considered.

4. Product water reclamation as an augment to imported irrigation water should befactored in to an economic assessment of membrane technology.

NEGATIVE

1. Little information is available at this time to assess the durability and life expectancyof NF membranes.

2. In addition, little information is available to assess the sensitivity of NF membranesto disinfectants.

3. Some research has been conducted on trace organics in feed water but not enoughto completely assess the impact of measured TOC levels on membraneperformance.

D. Solar Ponds

The use of salt-gradient solar ponds may play an important long-term role indrainage treatment for regions of the SJV. Solar ponds provide an environmentally safemethod for storage of agricultural drainage water salts and an alternative source ofenergy. A large quantity of salt is required to construct a solar pond. It takes theamount of salt in 146 acre-ft of 10,000 mg/L TDS drainage water to create one acre ofsolar pond with gradient and storage layers of 5 feet each. Agricultural drainage wateris concentrated 32 times for use in the storage zone of a solar pond.

Salt-gradient solar ponds are bodies of water in which the salinity at the bottom isgreater than the salinity at the surface. Depth may range from a fraction of a meter toseveral meters. Normally, as shown in the figure below, solar ponds consist of threezones: a relatively homogeneous and convective surface layer; a non-conductivegradient zone in which there are gradients in salinity and temperature; and ahomogeneous and convective bottom zone. The non-convective gradient zonetransfers heat by conduction only, water being opaque to thermal radiation at thetemperature of operation. It thus acts as a partially transparent window of low thermalconductivity that allows some solar radiation to penetrate and heat the bottom zone.The gradient remains gravitationally stable and non-convective despite, the heating ofthe bottom because the salinity gradient is sufficient to maintain a stabilizing densitygradient (Dickinson and Cheremisinoff, 1980).

10

The California Department of Water Resources, at its Demonstration DesaltingFacility in Los Banos, California, has investigated the use of a salt-gradient solar pondto treat agricultural drainage water from 1985 through 1989. The half-acre pondgenerated temperatures in excess of 180oF (82oC) (Engdahl, 1987). This investigationdemonstrated the use of the storage zone’s heat to generate electricity with aRankine-cycle engine. The use of the heat directly as the driving force for evaporationin a vertical tube evaporation process was also demonstrated (Kovac, 1990).

The solar pond should be viewed as an integrated system to reclaim orconcentrate agricultural drainage, store brine, and separate solid salts. Someprocesses being developed to treat and concentrate drainage water and separate saltswill be employed in this “solar pond” system. Work is needed to develop operatingguidelines, provide reliable estimates of potential energy production, and evaluatecritical economic factors. Additional work is needed to understand how these systemsprevent selenium in the water column from entering the food chain.

E. Chemical Reduction of Selenium with Zero-valent Iron

Iron filings (zero-valent iron) can be used to remove selenium from water. Theiron acts as both a catalyst and reductant (electron donor) for the reaction. Theselenium is reduced to selenite, Se(0), and selenide depending upon pH and O2 of thewater. Low pH and low oxygen favors the more reduced forms of selenium. In 1985,Harza Engineering Co. tested a pilot-scale processes using iron filings in flow-throughbeds. The testing was discontinued because the beds quickly cemented withprecipitates. The study did not conclusively identify the oxyhydroxide precipitates butrecent work suggests that “green-rust” (FeII

4FeIII2(OH)12SO4

.nH2O) may be the initialprecipitate, which is then oxidized to mangetite (Fe3O4) by nitrate and oxygen(Hansen, et al. 1996). Other likely precipitates include siderite (FeCO3) and ferrihydrite[Fe(OH)3]. The advantage of zero-valent iron is it can reduce the concentration of Se tovery low concentrations and might be useful as a polishing step following microbial

Gradient (Non-Convective)Surface (Convective)

Storage (Convective)

Salt-Gradient Solar Pond

Hot Brine Uses:Electricity

Evaporation

11

treatments. If the wastewater is anaerobic as a result of the microbial treatment, theformation of secondary precipitates is minimized.

F. Chemical Reduction of Selenium with Ferrous Hydroxides

A chemical reduction process for removal of selenate from subsurfaceagricultural drainage was developed and studied by the USBR Engineering andResearch Center, Denver, Colorado. Chemical reduction occurs between selenate andferrous hydroxide under alkaline conditions, producing iron oxides and elementalselenium, which are then removed by gravity settling.

USBR conducted bench level and batch reactors at their Denver laboratory andfound that selenate enriched synthetic drainage water was reduced to less than 5 µg/Lwithin one minute. These initial studies were followed in the field with a micro-pilot plantat Murietta Farms, Mendota, CA using SJV drainage water. Field studies indicatedslower rates of reactions. Although up to 90 percent of selenate was reduced, nearly6 hours reactor time was needed. It was found that dissolved oxygen, nitrate andbicarbonate interfered with the reduction process. Dissolved oxygen consumedexcessive ferrous hydroxide while nitrate and bicarbonate retarded the rate of reaction.Pretreatment of drainage water with sulfur dioxide can reduce dissolved oxygen andbicarbonate interference, but nitrate reduction was costly using chemical or physicalmethods. Without nitrate interference, drainage treatment by this technology isprojected to cost about $150 per acre-foot.

USBR has completed construction of a continuous flow reactor with pre-treatment devices, followed by ferrous hydroxide reduction. However, nitrate reductionis required for this treatment technology to be an economical process. USBR hasinvestigated innovative nitrate removal processes that are promising, but need furtherdevelopment. At present, USBR budget constraints and priorities have delayed furtherprogress on this technology.

The goal for selenium reduction is to achieve local receiving water standards(5µg/L) and this would require more than 99 percent reduction of the typical drainagewater. However, present basin management is directed to selenium load reduction.This shift in basin management approach would provide opportunity for processes thatare less costly in achieving load reduction, without meeting the stringent receiving waterstandards that are based on concentrations. Chemical reduction of selenate withferrous hydroxide in subsurface drainage can be achieved but will not meet the very lowstandard levels due to nitrate interference. Further studies are needed to control nitrateinterference at a reasonable cost.

G. Algal-Bacterial Selenium Removal Process

W.J. Oswald and his Applied Algae Research Group at the University ofCalifornia, Berkeley have developed a selenium treatment process consisting of aseries of specially-designed ponds. This group operated a 760-gallon per day prototypeAlgal-Bacterial Selenium Removal (ABSR) system near Mendota from 1986 to 1989.The best reductions in total soluble selenium were from 400 µg/L to less than 10 µg/L.Preliminary economic analyses conducted in 1988 indicated that ABSR facilities with

12

capacities of 3, 30, and 300 acre-feet per day would have a total costs of $272, $103,and $68 per acre-foot, respectively. The Mendota studies elucidated several keymicrobiological and physico-chemical mechanisms for selenium and nitrate removal andlaid the foundation for a current 10,000 to 20,000 gallon per day demonstration project(Lundquist et al., 1994).

ABSR Basics. The concept of the ABSR Process is to grow microalgae indrainage water at the expense of nitrate and then to utilize the naturally-settled algalbiomass as a carbon source for native bacteria which, in the absence of oxygen, reducethe remaining nitrate to nitrogen gas and reduce selenate to insoluble selenium. Theinsoluble selenium is then removed from the water by sedimentation in deep ponds and,as needed, by dissolved air flotation and sand filtration. Supplemental carbon sourcessuch as molasses can be employed as reductant in addition to algal biomass.

Past and current studies show a clear need to reduce dissolved oxygen andnitrate to low levels before selenate can be reduced. In the ABSR Process thisreduction is carried out by bacteria in Reduction Ponds (RPs) that are sufficiently deepand/or covered to exclude oxygen. Nitrate is also removed by assimilation into algalbiomass in High Rate Ponds (HRPs). Since nitrate concentrations in drainage waterare often as high as 90 mg/L as N compared to <0.5 mg/L for selenium, the carbonrequirement for nitrate reduction far exceeds that for selenium reduction. According tothe experience of Oswald’s group, the high sulfate concentration (2,000 to 4,000 mg/Las −2

4SO ) in drainage water does not appear to significantly interfere with nitrate orselenium reduction.

Selenium removed from the water column accumulates in settled algal-bacterialbiomass and inert materials in the bottom of the RPs. Because the biomass iscontinuously undergoing anaerobic decomposition, the volume of solid residuesincreases slowly. Removal and disposal of the solids in a landfill should not be requireduntil after many years of accumulation. Alternatively, the dried inert solids which containnitrogen, phosphorus, and selenium could be useful as a soil amendment and fertilizerin the eastern Central Valley where the soils are selenium deficient.

Large-scale ABSR Facilities are expected to pose much less hazard to wildlifethan the surrounding drainage channels or evaporation ponds. The concentration ofselenium in the shallow HRPs will be similar to or less than that of the drainagechannels themselves, and HRPs are continuously mixed by paddle wheels to preventsedimentation of particulates. Concentrated selenium will be sequestered in theanaerobic depths of the RPs. At 16- to 20-foot depth, the anaerobic RP floor cannotsupport invertebrate life and will not attract waterfowl.

Panoche Facility. The objective of the project hosted by the Panoche DrainageDistrict is to demonstrate the technical and economic feasibility of the ABSR Process forreduction of selenium loads from a portion of the combined drainage flows of severaldistricts or from selected drainage sumps within a single district.

The ABSR Facility consists of two identical ABSR systems treating a total of10,000 gallons per day as of October 1998. Different sequences of nitrate removal arebeing evaluated in each system. The Mode 1 system assimilates part of the nitrate in

13

HRP algae and denitrifies the rest using algae as the bacterial substrate, while theMode 2 system depends entirely on denitrification fueled by molasses and algalsubstrates.

Under Mode 1, the HRP has removed an averaged 50 mg/L NO3-N duringMarch to October 1998. Denitrification of the residual nitrate has been complete fromAugust to October 1998. Denitrification in the Mode 2 system has consistently reducednitrate-N from over 70 mg/L to less than 10 mg/L as N during April 1997 to October1998 except for February and March 1998 when impassable roads prevented RPsubstrate feeding.

Over 99 percent of the selenium in the drainage water influent is in the selenateform (SeO4

2-), but the raw pond effluents are a mixture of selenate, selenite (SeO32-),

and particulate selenium. As needed, particulate- and selenite-selenium can beremoved from the pond effluents with iron salt coagulation followed by clarification andfiltration. Selenate concentrations in the influent and effluents of the pond systems arecurrently used to determine selenium removal efficiencies.

The Mode 1 system reduced selenate concentration by 49 percent from March1998 through August 1998. The lowest concentrations were 39 µg/L selenate-Se and93 µg/L total soluble selenium. Since January 1997, the Mode 1 system removed37 percent of its cumulative selenate-Se mass load.

Selenate concentration dropped across the Mode 2 system by an average of83 percent from July 1997 through January 1998 and, after molasses feedingrecommenced, by 94 percent from April through August 1998. During the last periodeffluent selenate averaged 29 µg/L-Se with a low of 15 µg/L-Se. Total soluble seleniumaveraged 41 µg/L. Cumulative selenate-Se mass removal has been 86 percent sinceJanuary 1997.

The Mode 2 system has given predictable removals for over a year and will nowbe stress-tested by increased flow during the cooler winter months. During 1998,selenium and nitrate removal in the Mode 1 system has greatly improved, indicating thatthe cultures of algae and bacteria have matured. This system will continued to bemonitored under steady flow until removal rates stabilize.

Once stabilized, operating parameters such as flow rate, pond depth, andnutrient feed rates will be systematically adjusted to determine the conditions thatminimize the cost per mass of selenium removed. The main cost components to bereduced are supplemental carbon feed (molasses) and pond volume. Preliminaryresults indicate that full-scale ABSR facilities will require 4 to 6 acres per acre-ft/daytreated. The planned research efforts will provide better information on the costs ofselenium load reduction using full-scale ABSR facilities.

14

H. In Situ Treatment Utilizing Aquatic Algal Productivity

The term "in situ" refers to any open (instead of enclosed) on-site setup while"aquatic algal productivity" refers to any fully aquatic, light-driven biological process,where direct light input is required for part of the process. Thus, "algal" in this section isnot restricted to photosynthetic organism or to single-species processes. This broaddefinition was adopted in order to cover the widest range of current and pasttechnologies and research efforts in this area. In the SJV evaporation ponds, theprincipal purpose of remediation has been - and still is - alleviating the foodchainecotoxic effects of selenium in the ponds. Waterborne and sediment Sebioaccumulates readily into the aquatic biota (from primary producers to aquatic birdsand top predators) with concentration factors of 1,000 or higher (Ohlendorf, 1997; Maierand Knight, 1994). The extent of bioaccumulation depends on the route of exposure(e.g. diet, water, or sediment) and chemical form of Se (Besser et al., 1993; Maier andKnight, 1994). In Se-laden aquatic environments, chronic toxicity resulting from dietarySe uptake through the foodchain represents a far greater problem than acute toxicityassociated with direct water exposure (Maier and Knight, 1994). This chronic Setoxicity is, to an extent, related to the combination of waterborne Se concentration andSe bioaccumulation, as revealed at Kesterson Reservoir and other evaporation basinsof the San Joaquin Valley (Skorupa and Ohlendorf, 1991).

However, this relationship is not applicable to all aquatic environments and manycases have been reported where waterborne Se concentrations, Se bioaccumulation,and biological impacts did not correlate (e.g., Hamilton, 1997; Hamilton et al., 1997;Canton and Van Derveer, 1997; Lemly, 1993; Adams et al., 1997). Possibly alongsimilar lines in the SJV, a large decrease in Se content in avian eggs was recentlyobserved at the Rainbow Ranch evaporation pond (Kern County, CA) after a moderatedilution of waterborne Se concentration with agricultural tail water (Anthony Toto,CVRWQCB, Des Hayes, CDWR, and Joe Skorupa, USFWS, personal communication).

Consequently, there is a general consensus (Adams et al., 1997) that thecomplex Se biogeochemistry, particularly biotransformed Se forms in food organisms,may be the key to chronic Se effects expressed in species such as fish and birds.Since Se biogeochemistry varies with site conditions (e.g. fast-flowing river versusslow-flowing wetlands), Adams et al. (1997) expressed the need for site-specific waterquality criteria. Moreover, a departure from using waterborne concentrations (that is,using sediment-based water quality criteria) has also been proposed by Canton andVan Derveer (1997). Cautions regarding the indirect connection of waterborne Se toecotoxic effects were summarized specifically for Central Valley evaporation ponds byMaier and Knight (1992).

15

I. Algal Bioremediation Studies

In light of the emerging view that remediation involves complex, site-specific Sebiogeochemistry, algal bioremediation efforts in the SJV evaporation ponds appear tobe in their infancy. Although there are several algal studies with the stated goal ofselenium removal from water, it is clear from the preceding discussion that remediationinvolves far more than removal; the very definition of "remediation" requires that aprocess must reduce the factor(s) controlling the ecotoxic effects of selenium.

Recent algal studies at U.C. Davis, which have the long-term goal of alleviatingecotoxic effects, have begun to survey the selenium biogeochemistry ofnaturally-occurring algae in SJV evaporation ponds (Fan et al., 1997). Their goal is tounderstand the pathways and control points of ecotoxic effects in order to promote orreduce specific algal processes, as appropriate to foodchain accumulation of selenium.An additional advantage of keying on the naturally-occurring algae is that they arealready adapted to the physical-chemical conditions and variations of the SJVevaporation ponds.

Several species of microalgae isolated from SJV evaporation ponds transformselenium oxyanions into volatile alkylselenides, selenonium ions, and proteinaceousselenomethionine (Fan et al., 1997; Fan et al., in press). In particular, two cyanophytestrains isolated from the evaporation basins of the Tulare Lake Drainage District (TLDD)were very active in selenium volatilization. Up to 70 percent of the selenium wasdepleted from the medium containg 20-1,000 µg/L selenium by the algal activity and amajor fraction of the selenium loss was accounted for via the volatilization route. It ispossible that the algal selenium volatilization activity may be related to the recentobservation that waterborne selenium concentrations at one of the TLDD basins werenot increasing with rising salinity and that this trend was consistent year-around(Doug Davis, personal communication). The same observation was noted at KestersonReservoir by Saiki and Lowe (1987) and Schuler (1987). Selenium concentrationsgenerally were reduced from the San Luis Drain through a series of ponds, whereasTDS and boron increased through evapo-concentration. Additional research is neededto determine this relationship. Moreover, selenium volatilization and depletion fromwater were significantly enhanced by a sudden drop of the medium salinity. How thissalinity-induced change in selenium fate is related to the substantial drop in seleniumconcentration in avian eggs collected from the Rainbow Ranch evaporation pond isunclear. One possibility is that the enhancement of selenium volatilization induced bydecreased salinity may divert selenium away from bioaccumulation in aquatic birdsobserved at Rainbow Ranch. Further research will need to be conducted to unravel thispotentially important phenomenon.

Thus, it may be possible to further enhance this reduction of seleniumaccumulation into aquatic birds, if the mechanism(s) underlying the phenomenon isunderstood. If successful, such a "natural bioremediation" approach should represent areadily implementable and economical means for alleviating selenium contaminationproblems, not only in agricultural evaporation basins but also in other lentic systems(e.g. reservoirs and lakes) that receive Se-laden industrial waters.

16

Although selenium volatilization accounted for a major loss of total selenium fromwater in laboratory studies, a significant amount of selenium was also accumulated inthe algae, particularly in proteins where selenomethionine was the dominant form.Moreover, there was a major difference in selenium allocation into proteins amongdifferent algal species (Fan et al., in press); the significance of this finding is that thedifferent protein-bound selenium content may represent very different foodchain transferand ecotoxic potential among algae. This notion is supported by other studies,e.g., where diet with a higher level of protein led to a higher selenium content in fishtissues (Reidel et al., 1997). On the other hand, a higher protein content in water(e.g. from decaying algae) led to dramatically increased selenium volatilization inevaporation pond water by bacteria (Frankenberger and Thompson-Eagle, 1989), butthe ecotoxic consequences were not addressed in that study.

At this time, there appear to be no other algal studies that addressbioremediation per se in evaporation ponds. The above studies by Fan et al. still haveconsiderable work to be done because the factor(s) that control the ecotoxic foodchaineffects of selenium are not sufficiently understood. Consequently, it is not known whatpractical measurements must be made in order to assess the effectiveness of a processfor remediation purposes.

J. Selenium Removal from Water by Algal Processes

Studies of selenium removal from water - without addressing ecotoxicremediation - have shown interesting results. These are treated briefly here.

In 1986, Packer and Knight reported that the cyanobacteria ("blue-green algae")Synechoccus 6311 could be used to remove selenium from waters (Packer and Knight,1986). The mechanism of selenium uptake was proposed to be analogous to that ofS uptake, and the selenium was preferentially taken up over S at low sulfate levels,while the opposite was true at high sulfate levels. The remediation potential of theapproach was not discussed. This approach was apparently not pursued, possibly dueto the high sulfate salinity of evaporation ponds.

Over the last decade, a group based at UC-Berkeley have worked on analgal-bacteria system for selenium removal from water (Lundquist et al., 1994). Verybriefly stated, the system uses algae primarily as an organic carbon source for bacteria(Acinetobacter sp. and Pseudomonas sp.) to reduce the selenium oxyanions toinsoluble forms (e.g. colloidal Seo) that can be removed from water. In its present form(Benson et al., 1997), the focus of the system is on reducing waterborne selenium tomeet the discharge requirements into rivers, and does not appear to involve evaporationponds per se. The algal biotransformation activity was not reported and the remediationpotential for ecotoxic risk is unclear.

Nelson et al. (1997) at UC-Davis have developed a laboratory bacterial culture ofChromatium vinosum or Chlorobium limicola plus Desulfovibrio desulfuricans to reduceselenate to selenium-rich intracellular colloids probably containing Seo. The process is"algal" under our definition because the first two organisms are grown phototrophicallyunder anaerobic conditions, essential for control of the selenium chemistry. The

17

authors envision the process as useful in a bioreactor situation to remove seleniumoxyanions from the water. The remediation potential of the approach was notdiscussed.

K. Other Current Work that Includes Algal Processes

In addition to the efforts at algal selenium removal, there are other investigationscurrently underway that deliberately or incidentally involve algal processes. These aredescribed very briefly here.

Terry (1997) and Tanji (1997) describe a study using outdoor flow-throughwetlands for the removal of selenium from water. Because of the open design,including cells which have no vascular plants, the system is intentionally designed toinclude the effects of algae. As this system has only recently stablized, there are noclear conclusions that can be drawn at this time.

Parker et al. (1997) are using evaporation pond/wetland mesocosms to track thefate and partitioning of selenium. This is intended to be a more defined study of onedescribed by Terry (1997). Since this is a biogeochemical system, it will involve theaction of algae, although there is no intentional innoculation of any species.

L. Criteria for Success of In Situ Biologically Based Selenium TreatmentTechnology

Simple parameters such as waterborne selenium concentration are not likely tobe reliable indicators for success of in situ selenium bioremediation technologies.The main reasons for this are reiterated here: 1) The goal of selenium remediation is tominimize ecotoxic risk, not just removing selenium from waters; 2) selenium exposureand toxicity in birds and fish primarily result from selenium biotransformations andtransfer pathways through the foodweb, which are not simple functions of waterborneselenium concentration. Waterborne selenium concentrations much below 5 µg/L (thecurrent EPA selenium freshwater quality criterion) have been shown to have adverseeffects on fish (e.g. Lemly, 1993). Conversely, there have been cases where noapparent selenium toxicity was observed with waterborne selenium concentrationsmuch higher than 5 µg/L (e.g. Canton and Van Derveer, 1997; Hamilton et al., 1997).

The consensus conclusion, based on such results, was that waterborne seleniumconcentrations alone would be inadequate in assessing the selenium impact on theaquatic community (from a 9-member expert panel in a recent "Peer ConsultationWorkshop on Selenium Aquatic Toxicity and Bioaccumulation" held by theUS Environmental Protection Agency, May, 1998). It was clear from the paneldiscussion that more reliable parameters such as selenium burden in representativetissues and tissue compartments (e.g. proteins) will be needed for selenium impactassessment. In light of these new findings in the last decade, EPA is initiating theprocess of re-evaluating the current chronic water quality criterion for selenium. Morereliable parameters will lead to better regulatory criteria for protecting the aquaticcommunity, and will be equally important for assessing the efficacy of any mitigationmeasure.

18

M. Volatilization

Volatilization of selenium, a naturally occurring component of selenium cycling inthe biosphere, is a primary mechanism of dissipating various selenium ions out of waterand sediments. Comprehensive reviews have been published on bioremediationtechnologies to deselenify soils, sediments and water of selenium by volatilization(Frankenberger and Karlson, 1991; Thompson-Eagle et al. 1991b; Frankenberger andKarlson, 1992; Thompson-Eagle and Frankenberger, 1992; Karlson and Frankenberger,1993; Frankenberger and Karlson, 1994a; Frankenberger and Losi, 1995; and Losi andFrankenberger, 1997). Management factors affecting volatilization of selenium fromsediments and water are summarized in Tables 1 and 2.

Frankenberger and his co-workers have conducted pioneer research incharacterizing volatilization of selenium as a bacterial and fungal reaction. Thistransmethylation reaction involves homocysteine, methionine, cobalamin derivativesand reducing agents such as thiols including glutathione. Methylating microorganismshave been isolated in many sediments and evaporation pond waters throughout theSJV. The addition of protein sources as well as pectin can stimulate volatilization up to300-fold of the native natural rate (Thompson-Eagle and Frankenberger, 1990). Theprimary product of selenium volatilization is dimethylselenide (DMSe) which is anontoxic species as documented in acute toxicity studies. Selenium removal rates havebeen reported as high as 38 percent for protein-treated pond waters after140 days. Volatilization can be inhibited at very high nitrate concentrations but sulfatehas no effect on DMSe release. Laboratory studies with 75Se-labelling were employedto identify optimum management practices to accelerate selenium volatilization. Thisbioremediation technology is highly dependent on specific carbon amendments (pectinand proteins), pH, temperature, moisture, aeration, and activators (co-factors).Selenium biomethylation is protein/peptide-limited rather than nitrogen-, amino-, orcarbon-limited. Crude casein and its components,δ-, β-, and κ-caseins, and peptidesare equally stimulatory producing greater than 50-fold enhancement in DMSe yield.

To demonstrate the viability of volatilization of selenium as an effectivebioremediation technology to dissipate selenium from soil and sediments,Frankenberger (unpublished) has conducted an experiment where the San Luis Drainsediment was collected, homogenized and placed in mesocosms in the greenhouse.These mesocosms consisted of contained systems opened to the atmosphere with thesediment 6 inches deep. The treatments consisted of moist-only, application of groundorange peel and a protein mixture including casein and albumin. The containedsediments were mixed and stirred to promote aeration on an interval of twice a week.Figure 1 shows the decline in total selenium in the sediments on monthly intervals. It isevident from these data that orange peel and the protein mixture are more effective inpromoting selenium volatilization. After 10 months, approximately 13.6 percent of theselenium was removed in the moist treatment compared to 37.0 percent upon theaddition of orange peel and 45.8 percent when treated with the protein mixture. Thisstudy unequivocally demonstrates that selenium volatilization is a viable technology toremove selenium from soils and sediments. This desalinification procedure has beenoutlined in detail by a patent filed by Frankenberger et al. (1989).

19

20

Higher plants also possess the capacity to volatilize selenium (Lewis, 1971;Terry et al., 1992; Biggar and Jayaweera, 1993). This process occurs in both selenium-accumulator and non-selenium-accumulator plant species. Selenium non-accumulatorstypically release DMSe, while accumulators mainly volatilize DMDSe (Evans et al.,1968; Lewis, 1971). Generally, plant species differ substantially in their ability tovolatilize selenium (Terry et al., 1992). Indian mustard and other Brassica species werefound to be particularly effective volatilizers of selenium. There was evidence that theability to volatilize selenium was associated with the ability to accumulate selenium inplant tissues (Terry et al., 1992). Recently, it has been shown that roots are the mainsites of selenium volatilization in plants (Zayed and Terry, 1994). This is especiallyinteresting since plant roots can be used to filter out selenium from polluted watersthrough uptake and accumulation in plant tissues in addition to volatilization. In fact,rhizofiltration, the use of plant roots to remove pollutants from water, is an emergingenvironmental cleanup technology, not only for selenium but also for many other traceelements and heavy metals (Dushenkov et al., 1995). Roots of many hydroponicallygrown terrestrial plants were shown to be effective in removing pollutants from waterdue to uptake and concentration in their tissues (Dushenkov et al., 1995). In the case ofmetalloids such as Se, volatilization provides an additional pathway of removal, whichaccelerates the rate of pollutant removal from water.

N. Biological Precipitation

There are number of reports on microorganisms which can reduceselenooxyanions into elemental selenium. However, in most cases, these bacteriaexhibit low growth rates and require strict anaerobicity. Reduction may be inhibited byalternative electron acceptors such as −

3NO which is common in seleniferous wastestreams.

Recent work has suggested that certain facultative organisms are able to reduce−24SeO under anaerobic or microaerophilic conditions. A recently isolated −2

4SeO reducingorganism, Thauera selenatis is reportedly a facultative anaerobe that reduces −2

4SeOonly while growing anaerobically (Macy, 1994). Losi and Frankenberger (1997b) haveisolated a bacterium from the San Luis Drain (San Joaquin Valley, CA) identified asEnterobacter cloacae SLD1a-1 which is a facultative anaerobe and is capable of using

−23NO and −2

4SeO as terminal electron acceptors during anaerobic growth. Nitrate doesnot seem to interfere with −2

4SeO reduction by this organism.

Washed-cell suspension experiments with Enterobacter cloacae SLD1a-1revealed that selenite is a transitory intermediate in reduction of −2

4SeO being producedand rapidly reduced concomitantly. Nitrate is also reduced concomitantly and at a muchhigher rate than −2

4SeO . Although this reaction is enzymatic, reduction of eitheroxyanion does not appear to be an inducible process. Recent studies have shownoptimum conditions (including pH, temperature, and electron donors) for −2

4SeOreduction and removal of selenium from drainage water by Enterobacter cloacaeSLD1a-1 (Losi and Frankenberger, 1997b,c).

21

Enterobacter cloacae SLD1a-1 is capable of reducing selenate and selenite intodimethylselenide as well as precipitating the soluble oxyanions into an insolubleelemental selenium. This bacterium is active in volatilizing selenium under aerobicconditions and precipitates selenium into Seo under anaerobic conditions. Thus underboth modes of respiration (aerobic and anaerobic) it has two separate mechanisms forremoving selenium out of its surrounding environment. Optimum managementschemes can be manipulated to enhance either process in the removal of seleniumfrom drainage water.

Several treatment technologies have been investigated that utilize seleniumbioreduction to remove selenium oxyanions from solution with a goal of treatingagricultural drainage water. An extensive review of these technologies has recentlybeen presented by Losi and Frankenberger (1997a). Selenium present in drainagewaters primarily exists in the soluble oxidized selenate form. By supplying an organiccarbon source as an electron donor, selenium can be reduced to selenite and then toelemental selenium. Elemental selenium is a particulate and can be removed byconventional solid separation processes. This can be illustrated by the followingpseudo reaction:

Se (+ VI) + Bacteria + Organic CarbonSelenate(soluble)

→ Se(+IV)Selenite(soluble)

→ Se(0)Elemental Se

(particle)

Because nitrate is also present in the drainage water, it is reduced throughbacterial denitrification. Oxygen must be excluded from the reactor or reduction ofnitrate and selenium will not occur. Because nitrate is present in much higherconcentrations than selenium, most of the organic carbon is used for nitrate reduction.Reactor configuration influences the efficiency and speed of the conversion and sodifferent types of reactors have been investigated.

Several pilot-scale projects that use anaerobic reduction reactions to precipitateselenium from drainage water have been tested. These treatment processes varybased on reactor design and the type of carbon source and nutrients provided to themicroorganisms. Reactor designs include above ground tanks containing sludge beds,fluidized beds, or fixed films. The carbon sources that have been tested includemethanol, Steffens waste from sugarbeet processing, acetate from acetic acid, andmolasses. The most significant research and demonstration of selenium removal bybiological processes in reactor systems has been conducted by EPOC AG at MuriettaFarms, Dr. Joan Macy, University of California, Davis, and Engineering ResearchInstitute (ERI) at California State University, Fresno, CA. In studies conducted by thoselisted above, the following conclusions can be reached.

22

1. Soluble selenium in drainage water can be reduced from 300 or 500 ppb to30 ppb or less.

2. A two-stage reactor system with a total hydraulic detention time on the order of afew hours appears to be an effective combination. The upflow anaerobic sludgeblanket reactor (UASBR) followed by a fluidized bed reactor is particularly well-suited to treating drainage water because of this combination’s ability to handleinorganic precipitation.

3. The total cost for one million gallons per day would be estimated at $440 to $520acre-foot of water treated, with operating costs representing $340 to $420acre-foot. The cost for the organic carbon that is added to the water accounts for25 to 5 percent of the operating costs if methanol is used.

The advantages of the biological precipitation reaction are (i) soluble seleniumcan be reduced by up to 95 percent, (ii) low production of solid residuals for disposaldue to slow anaerobic growth rates, (iii) low required detention time allows use ofstandard engineering sizes of treatment systems, and (iv) the system is enclosed withno exposure to the environment or wildlife. The major disadvantage of this technologyis that soluble selenium residual of 30 ppb or greater is likely from biological treatmentonly and organic carbon costs are high unless low cost alternatives can be used.

O. Flow-Through Wetlands

Constructed wetlands constitute a complex ecosystem, the biological andphysical components of which interact to provide a mechanical and biogeochemicalfilter capable of removing many different types of contaminants from water. They havebeen used to cleanup municipal wastewater, stormwater runoff, and many other typesof polluted wastewaters in the USA and in Europe (Kadlec and Knight, 1996).Constructed wetlands are orders of magnitude lower in cost than other treatmentsystems; however, the science of understanding wetland selenium detoxificationmechanisms is in its infancy.

There is evidence that significant losses of Se, through biologically mediatedmethylation and volatilization, can occur in wetlands. The toxicity of the volatileselenium compounds, such as dimethylselenide (DMSe), are 600-700 times lower thanthe inorganic forms (Karlson and Frankenberger, 1988). Finding the conditions that areoptimum for volatilization has been one of the goals of constructed-wetland studies.Concerns over low volatilization rates of DMSe from aquatic systems and the gradualaccumulation of selenium in the sediments and vegetation may limit this method. Theattractiveness of this method is the low cost and large volumes of water that could betreated. The biggest concern with wetland treatment systems is the potential use bywaterfowl and the similarity to the conditions that existed at Kesterson Reservoir.

One indication that wetlands might be useful in the removal of selenium fromwastewaters came from a study of a 36-hectare constructed wetland located adjacent tothe San Francisco Bay, California. Some Bay area oil refineries have discharged intothe Bay wastewater containing selenite at levels that substantially exceed thosepermitted under the guidelines established by both state and federal regulators.

23

Chevron has supported research directed by Norman Terry at UCB to determine thefeasibility of wetlands in removing selenium from their wastewaters. Analysis of thewetland inlet and outlet waters showed that the constructed wetland was successful inremoving 89 percent of the selenium from the wastewater passing through it. Inflowselenium concentrations of 20-30 µg L-1 decreased to <5 µg L-1 in the outflow(Hansen et al., 1998). The fact that the Chevron wetland was processing up to10 million liters of oil refinery effluent each day strengthens the significance of the roleof constructed wetlands in removing selenium from large volumes of wastewater.

Constructed wetlands remove selenium by reduction to insoluble forms which aredeposited in the sediments, by accumulation into plant tissues, and by volatilization tothe atmosphere, i.e., through biological volatilization of plants, plant/microbeassociations and microbes alone. Cooke and Bruland (1987) estimated that as much as30 percent of selenium introduced into the ponds of Kesterson Reservoir in the SJV ofCalifornia was lost by volatilization into the atmosphere through biomethylation and thatthe process varied seasonally. Allen (1991) tested the ability of a constructed wetlandto remove selenium from water spiked with selenite. She found that the rate ofselenium removal by the wetland exceeded 90 percent of the total selenium introducedand considered that biological volatilization was an important component of this loss.Recently, Zhang and Moore (1997) examined the role of volatilization in the removal ofselenium by a wetland system, using wetland microcosms spiked with different forms ofselenium. Their results indicate that natural selenium volatilization is an importantprocess removing selenium from wetland systems. Sediment and plants were the majorproducers of volatile selenium from the system. Actual field measurements of seleniumvolatilization performed by N. Terry and his co-workers have shown that biologicalvolatilization might have accounted for as much as 10-30 percent of the total seleniumremoved from the San Francisco Bay constructed wetland (Hansen et al., 1998). Theymeasured rates of selenium volatilization as high as 330 µg m-2 soil surface day-1 fromareas vegetated with rabbitfoot grass. Thus, it appears that biological volatilization ofselenium is a significant pathway of selenium removal in wetlands.

The success of the San Francisco Bay constructed wetland in removing seleniuminspired a group of scientists under the auspices of the University of CaliforniaSalinity/Drainage Task Force to initiate a flow-through wetland at the Tulare LakeDrainage District site in the spring of 1996. The purpose of this trial was to test theeffectiveness of the flow-through wetland in removing selenium from selenium-ladenagricultural drainage waters before discharge into the evaporation basins. Tenrectangular wetland cells, each 250 x 50 ft in dimension, were constructed. Thewetland cells were then flooded and planted with various types of vegetation such thateach cell contained one or more of the following plant species: salt marsh bulrush, balticrush, cattail, cord grass, rabbitfoot grass, saltgrass, tule, and widgeon grass. One cellwas left without plants as a control natural cell. By the summer of 1997, all except theopen water cells were established with vascular plants. These wetland cells arecurrently being monitored in order to evaluate their performance in selenium removalfrom agricultural drainage waters containing relatively low levels of selenium(20-30 µg L-1). Surface water, sediment pore water, plant tissue, and sediment samplesare being collected monthly to determine the removal efficiency, fate of selenium, andthe oxidation state of water soluble selenium (Terry, 1997). Plexiglass chambers thatcover whole plants are being used to monitor selenium volatilization rates. Preliminary

24

data suggest that 21 percent to 89 percent of the incoming selenium is being removedfrom the drainage water. Work is ongoing to determine the relative partitioning ofselenium into sediments, plant biomass, and gaseous losses. To evaluate seleniumtransfer through the food chain, analysis of selenium load in primary producers andrepresentative invertebrate species is underway. The most recent results show that:i) the wetland cells are capable of significantly reducing the concentrations of seleniumin drainage water, ii) the most efficient cell in reducing selenium was that planted withcattail and widgeon grass, and iii) the efficiency of the wetland in removing seleniumappears to be increasing from 1997 to 1998.

In conjunction with this field-scale project, a mesocosm-scale project is ongoingat U.C. Riverside. This project is designed to address some of the shortfalls of the field-scale project which include non-replication and an inability to measure leaching losses.The mesocosm wetlands consist of polyethylene drum-halves linked in series and filledwith 30 cm of soil. Three drum-halves are linked in a cascade design to represent thebeginning, middle, and end of a long, narrow wetland cell. Twelve mesocosm seriesallow for 4 treatments with 3 replications. In the first part of the study, the treatmentswere cattails, rabbitsfoot grass, a shallow water control, and a deep water control(no vegetation). Preliminary analysis of the influent and effluent indicates that thecumulative selenium load was reduced by 42 percent in the vegetative treatments and28 percent in the non-vegetative treatments. No significant difference in total seleniumreduction was observed between the two vegetation types (Parker et al., 1997).Preliminary results on the non-vegetative treatments suggest that about half of the losscould be attributed to volatilization.

P. Subsurface Flow Wetlands

Constructed subsurface flow (SSF) treatment wetlands can also be used toremove selenium from irrigation drainage water. These treatment wetlands systemswould be especially effective where selenium concentrations and loadings are high andflow volumes are low. The wetland would consist of gravel beds that serve as therooting medium for salt-tolerant emergent wetland plants (such as bulrush), with theagricultural drainage water flowing through the gravel (Fig. 2). As the water flowsthrough the gravel bed, oxidation-reduction processes in the plant root zone (especiallydue to the interactions of the microbial community and the plant roots) would removeselenium through a combination of uptake and volatilization. These processes are wellunderstood for treatment of conventional pollutants in wastewater (Kadlec and Knight,1996), and the same principles, coupled with knowledge about the oxidation-reductioncycling and volatilization of selenium (e.g., Frankenberger and Karlson 1994b,Oremland 1994, Terry and Zayed 1994), could be applied to treatment of selenium inagricultural drainage water.

25

26

Besides reducing selenium in drainage water to safe levels, another advantage ofthis technology would be the development of a treatment system that can beconstructed within individual sumps or agricultural drains so the loss of arable landwould be minimized. The primary advantage of subsurface flow treatment wetlands arethat they reduce the potential for an aquatic food chain pathway for selenium exposureto fish or birds.

The SSF treatment system will optimize volatilization through microbial-mediatedreactions as the primary removal mechanism. This technology could also evaluate thepotential of a specific bacterium, Enterobacter cloacae SLD1a-1, that has been shownto be effective in removing selenium from agricultural drainwater with little or noinhibitory effect of nitrate. Thus, the use of a replaceable, selenium-adsorptive or -precipitative treatment medium could be evaluated in comparison to (or integration with)a volatilization-based removal system.

III. CONCLUSIONS

This review evaluates the most recent findings on treatment technologies inremoval of selenium from agricultural drainage water. New recent advances have beenmade in many technologies since the Agricultural Drainage Treatment Technologyreview was released in July 1998 by the San Joaquin Valley Drainage Program. Theenvironmental impact of selenium and other trace elements in the SJV drainage waterwas not fully recognized until about 1985, when high levels of selenium were identifiedin biota in Kesterson Reservoir. However, long before the selenium problem emerged,drainage water reclamation was being seriously considered at the tubular reverseosmosis plant in Firebaugh in the early 1970s. The motivation for construction of thisexperimental facility arose from two fundamental issues relevant to SJV agriculturaldrainage. Of primary concern was augmentation of irrigation water supplies bydrainage water desalinization. This was indeed a challenging application of theemerging technology under development at the time. A second goal was directedtowards reduction of drainage water volume. Management of salt accumulation couldthen be enhanced by such waste minimization technology.

These goals are especially relevant nearly 30 years later. State-of-the-artmembrane technology has developed to a point where economical desalinization ofdrainage water should be considered. In addition, solar pond technology, the potentialwhich has not been adequately explored, should be seriously considered from aperspective of salt management as well as energy production. These technologiescertainly play important roles in desalinization and salinity control in future managementof drainage water. Recently, U.S. Filter Corporation has demonstrated a microfiltrationtechnology in testing its ability to remove selenium and nitrates from the Alamo River inthe Imperial Valley, with the primary goal of recycling the agricultural drainage asirrigation water equivalent to the quality of the Colorado River. Although many of thesephysical technologies are quite promising in removing selenium and other toxic ionsfrom the water, their cost is prohibitive on a large scale application.

27

Chemical reduction of selenium with zero valent iron and ferrous oxides is lesseffective in water of high redox potential.

Biological treatment holds the most promise because of its cost effectiveness andpermanent removal of selenium from the matrix. Bacteria, plants and micro-algaeactive in selenium precipitation and volatilization are the most promising players inremediation of drainage water. Previous studies have shown that in aquatic systemsthe major mechanism of removal of selenium is through reductive precipitation. Themajor problem with volatilization of selenium, particularly from water, is that themethylated selenium species are water-soluble and thus their release into theatmosphere is subject to entrapment.

In reviewing each of these treatment technologies, it is imperative to note thatexciting advances are being made on a short-term investment. These biological systemsserve as biogeochemical filters capable of removing selenium through assimilation,bacterial precipitation, and volatilization. Technologies should be promoted that arecost-effective and pose the least threat to the environment with safety to birds and otherwildlife. Mass balances should be very carefully calculated to understand the mechanismof selenium removal. With intense research in treatment technologies, perhaps acombination of biological, chemical and physical techniques may be useful indevelopment of a cost-effective strategy in treatment of agricultural drainage water.

28

REFERENCES

Adams, W. J., K. V. Brix, A. Fairbrother, and R. Cardwell. 1997. Understandingselenium in the aquatic environment: A Learned Discourse.SETAC News 17: 22-23.

Allen, K. N. 1991. Seasonal variation of selenium in outdoor experimental stream-wetland systems. J. Environ. Qual. 20:865-868.

Atkinson, R., S.M. Aschmann, D. Hasegawa, E.T. Thompson-Eagle and W.T.Frankenberger, Jr. 1990. Kinetics of the atmospherically important reactions ofdimethyl selenide. Environ. Sci. Technol. 24:1326-1332.

Ayers, H.S. and D.W. Westcot. 1985. “Water Quality and Agriculture,” Food andAgricultural Organization of the United Nations—Irrigation and Drainage paperNo. 29, Rev. 1. Rome.

Benson, S., W. J. Oswald, and T. Leighton. 1997. Demonstration of the Algal-BacterialSelenium Removal Technology for Irrigation Drainage Water Treatment.Handout from and presentation at the 1997 Salinity/Drainage Annual Meeting,Sacramento, CA.

Bergman, R. A. 1996. Cost of membrane softening in Florida. J. Amer. Water WorksAssoc. 88:32-43.

Besser, J. M., T. J. Canfield, and T. W. La Point. 1993. Bioaccumulation of organic andinorganic selenium in a laboratory food chain. Environ. Toxicol.Chem. 12: 57-72.

Biggar, J.W. and G.R. Jayaweera. 1993. Measurements of selenium volatilization in thefield. Soil Sci. 155:31-36.

Calderone, S.J., W.T. Frankenberger, Jr., D.R. Parker and U. Karlson. 1990. Influenceof temperature and organic amendments on the mobilization of selenium insediments. Soil Biol. Biochem. 22:615-620.

Canton, S. P. and W. D. Van Derveer. 1997. Selenium toxicity to aquatic life: Anargument for sediment-based water quality criteria. Environ. Toxicol.Chem. 16: 1255-1259.

Cherr, G. N. and R. M. Higashi. 1997. Toxicity Identification Evaluations are inherentlyresearch projects: A Learned Discourse. SETAC News 17: 19-20.

Cooke, T.C. and K.W. Bruland. 1987. Aquatic chemistry of selenium: Evidence ofbiomethylation. Environ. Sci.Technol. 21:1214-1219.

Dickenson, W. C. and P. N. Cheremisinoff, eds. 1980. Solar Energy TechnologyHandbook. Marcel Dekker, Inc. New York, NY. pp. 345-346.

29

Dushenkov, V, P.B.A. N. Kumar, H. Motto and I. Raskin. 1995. Rhizofilteration: the useof plants to remove heavy metals from aqueous streams. Environ. Sci.Technol. 29:1239-1245.

Engdahl, D. D. 1987. Technical Information Record on the Salt-Gradient Solar PondSystem at the Los Banos Demonstration Desalting Facility, CaliforniaDepartment of Water Resources, San Joaquin District. Page 1.

Evans, C.S., C.J. Asher and C.M. Johnson. 1968. Isolation of dimethyl diselenide andother volatile selenium compounds from Astragalus racemosus (Pursh.).Aust. J. Biol. Sci. 21:13-20.

Fan, T. W.-M. and R. M. Higashi. 1998. Biochemical fate of selenium in microphytes:Natural bioremediation by volatilization and sedimentation in aquaticenvironments. In W.T. Frankenberger and R.A. Engberg (eds.), EnvironmentalChemistry of Selenium. Marcel Dekker, Inc., New York, pp. 545-563.

Fan, T. W.-M., A. N. Lane, and R. M. Higashi. 1997. Selenium biotransformations by aeuryhaline microalga isolated from a saline evaporation pond. Environ. Sci.Technol. 31:569-576.

Fan, T. W.-M., A. N. Lane, D. Martens, and R. M. Higashi. Synthesis and structurecharacterization of selenium metabolites, Analyst. In press.

Flury, M., W.T. Frankenberger, Jr. and W.A. Jury. 1997. Long-term depletion ofselenium from Kesterson dewatered sediments. Sci. Total Environ. 198:259-270.

Frankenberger, Jr., W.T. and S. Benson (eds.). 1994. Selenium in the Environment.Marcel Dekker, New York, NY. pp. 456.

Frankenberger, Jr., W.T. and R. A. Engberg (eds.). 1998. Environmental Chemistry ofSelenium. Marcel Dekker, New York, NY. pp. 713.

Frankenberger, Jr., W.T. and M.E. Losi. 1995. Applications of bioremediation in thecleanup of heavy metals and metalloids. pp. 173-210. In Bioremediation:Science and Applications. SSSA Special Publication 43. Soil Sci. Soc. Amer.,Madison, WI.

Frankenberger, Jr., W.T. and U. Karlson. 1988. Dissipation of soil selenium bymicrobial volatilization at Kesterson Reservoir. U.S. Dept. of the Interior, Bureauof Reclamation, Contract No. 7-FC-20-05240.

Frankenberger, W.T., Jr. and U. Karlson. 1989. Environmental factors affectingmicrobial production of dimethylselenide in a selenium-contaminated sediment.Soil Sci. Soc. Am. J. 53:1435-1442.

Frankenberger, Jr., W.T. and U. Karlson. 1991. Bioremediation of seleniferous soils.pp. 239-254. In R. E. Hinchee and R. F. Olfenbuttel (eds.), On-SiteBioreclamation. Butterworth-Heinemann, Stoneham, MA.

30

Frankenberger, Jr., W.T. and U. Karlson. 1992. Dissipation of soil selenium bymicrobial volatilization. pp. 365-381. In D. C. Adriano (ed.), Biogeochemistry ofTrace Metals. Lewis Publ., Boca Raton, Florida.

Frankenberger, Jr., W.T. and U. Karlson. 1994a. Campaigning for bioremediation.Chemtech 24:45-51.

Frankenberger, Jr., W.T. and U. Karlson. 1994b. Soil management factors affectingvolatilization of selenium from dewatered sediments.Geomicrobiol. J. 12:265-278.

Frankenberger, W.T., Jr. and U. Karlson. 1995. Volatilization of selenium from adewatered seleniferous sediment: A field study.J. Indust. Microbiol. 14:226-232.

Frankenberger, W. T. and E. T. Thompson-Eagle. 1989. In Situ Volatilization ofSelenium. II. Evaporation Ponds. Final Report to the San Joaquin ValleyDrainage Program, U.S. Bureau of Reclamation contract #7-FC-20-05110,pp. 1-28.

Hamilton, S. J. 1997. Evaluation of contaminant impacts on razorback sucker inrestored floodplain habitats and remediation techniques. Presentation at theSymposium "Understanding Selenium in the Aquatic Environment",Salt Lake City, UT, March 1997.

Hamilton, S. J., K. J. Buhl, F. A. Bullard, and S. F. McDonald. 1997 . On-site toxicitytest with razorback suckers at Ouray NWR, Utah. Presentation at theSymposium "Understanding Selenium in the Aquatic Environment",Salt Lake City, UT, March 1997.

Hanna, Jr., G. P., J. A. Kipps, L. P. Owens. 1990. Agricultural drainage treatmenttechnology review. Memorandum Report to the San Joaquin Valley DrainageProgram. Sacramento, CA. Under the U.S. Bureau of ReclamationOrder No. 0-PG-20-01500.

Hansen, H.C.B., C. B. Koch, H. Nancke-Krogh, O. K. Borggaard and J. Sorensen. 1996.Abiotic nitrate reduction to ammonium: key role of green rust. Environ. Sci.Technol. 30:2053-2056.

Hansen, D., P.J. Duda, A. Zayed and N. Terry. 1998. Selenium removal by constructedwetlands: Role of biological volatilization. Environ. Sci. Technol. 32:591-597.

Hothem, R. L. and H. M. Ohlendorf. 1989. Contaminants in foods of aquatic birds atKesterson Reservoir, California, 1985. Arch. Environ. Contam.Toxicol. 18:773-786.

Jacobs, L. W. (ed.). 1989. Selenium in Agriculture and the Environment. Soil ScienceSociety of America Special Publication Number 23. pp. 233.

31

Kadlec, R. H. and R. L. Knight. 1996. Treatment Wetlands. Lewis Publishers,Boca Raton, FL.

Karlson, U. and W.T. Frankenberger, Jr. 1988. Effects of carbon and trace elementaddition on alkyselenide production by soil. Soil Sci. Soc. Am. J. 52:1640-1644.

Karlson, U. and W.T. Frankenberger, Jr. 1989. Accelerated rates of seleniumvolatilization from California soils. Soil Sci. Soc. Am. J. 53:749-753.

Karlson, U. and W.T. Frankenberger, Jr. 1990. Volatilization of selenium fromagricultural evaporation pond sediments. Sci.Total Environ. 92:41-54.

Karlson, U. and W.T. Frankenberger, Jr. 1993. Biological alkylation of selenium andtellurium. pp. 185-227. In H. Sigel and A. Sigel (eds.), Metal Ions in BiologicalSystems. Marcel Dekker, Inc., New York, N.Y.

Karlson, U., W.T. Frankenberger, Jr. and W.F. Spencer. 1994. Physiocochemicalproperties of dimethyl selenide and dimethyl diselenide. J. Chem. Eng.Data 39:608-610.

Kovac, K., D. W. J. Hayes and H. H. Sephton. 1990. Brine concentration utilizing solarpond heat with a vertical-tube foamy evaporator, Biennial ConferenceProceedings, National Water Improvement Supply Association. Pages 19-48.

Le Gouellec, T., M. Elimelech and J. Glater. 1997. High performance nanofiltrationmembranes for agricultural wastewater Reclamation. Proc. IDA World Congresson Desalination and Water Reuse, Madrid, Spain. 97-111.

Lemly, A. D. 1993. Teratogenic effects of selenium in natural populations of freshwaterfish. Ecotoxicol. Environ. Saf. 26, 181-204.

Lewis B.G. 1971. Volatile selenium in higher plants: the production of dimethyl selenidein cabbage leaves by the enzymatic cleavage of methylselenomethionineselenonium salt. Ph. D. thesis, University of California, Berkeley.

Loeb, S. and S. Sourirajan. 1960. Sea water demineralization by means of asemipermeable membrane. UCLA Dept. of Engineering Report 60-60.

Losi, M.E. and W.T. Frankenberger, Jr. 1997a. Bioremediation of selenium in soil andwater. Soil Sci. 162:692-702.

Losi, M.E. and W. T. Frankenberger, Jr. 1997b. Reduction of selenium oxyanions byEnterobacter cloacae SLD1a-1: Isolation and growth of the bacterium and itsexpulsion of selenium particles. Appl. Environ. Microbiol. 63:3079-3084.

Losi, M.E. and W. T. Frankenberger, Jr. 1997c. Reduction of selenium oxyanions byEnterobacter cloacae strain SLD1a-1: Reduction of selenate to selenite. Environ.Toxicol. Chem. 16:1851-1858.

32

Lundquist, T. J., F. B. Green, R. B. Tresan, R. D. Newman, W. J. Oswald and M. B.Gerhardt. 1994. The algal bacterial selenium removal system: Mechanisms andfield study. pp. 251-278. In W. T. Frankenberger, Jr. and S. Benson, (Ed.),Selenium in the Environment. Marcel Dekker, Inc., New York, N.Y.

Macy, J.M. 1994. Biochemistry of selenium metabolism by Thauera selenatis gen. nov.sp. nov. and uses of the organism for bioremediation of selenium oxyanions inSan Joaquin Valley drainage water. pp. 421-444. In W. T. Frankenberger, Jr. andS. Benson, (Ed.), Selenium in the Environment. Marcel Dekker, Inc.,New York, N.Y.

Maier, K.J. and A. W. Knight. 1992. Selenium in Central Valley Evaporation PondSystems. Final Technical Report to California Department of Water Resources,contracts #B-57232 and #B-57232-1. Land, Air and Water Resources PaperNo. 100020, University of California, Davis, pp. 73-81.

Maier, K.J. and A. W. Knight. 1994. Ecotoxicology of selenium in freshwater systems.Rev. Environ. Contam. Toxicol. 134: 31-48.

McCutchan, J. W., V. Goel, D. B. Bryce, and K. Yamamoto. 1976. Reclamation of fielddrainage water. Desalination 19:153-160.

Nelson, D. C., W. H. Casey, J. D. Sisson, E. E. Mack, A. Ahmad, and J. S. Pollack.1996. Selenium uptake by sulfur-accumulating bacteria. Geochim. Cosmochim.Acta 60: 3531-3539.

Ohlendorf, H.M. 1989. Bioaccumulation and effects of selenium in wildlife. pp. 133-177.In L.W. Jacobs (Ed.), Selenium in Agriculture and the Environment. SSSA Spec.Publ. 23. American Society of Agronomy and Soil Science Society of America,Madison, WI.

Ohlendorf, H. M., G.M. Santolo 1994. Kesterson Reservoir--past, present, and future:An ecological risk assessment. pp 69-117. In W. T. Frankenberger, Jr. and S.Benson, (Eds.). Selenium in the Environment. Marcel Dekker, Inc.,New York, N.Y.

Ohlendorf, H. 1997. Toxicological effects of selenium on birds. Pres. Symposium"Understanding Selenium in the Aquatic Environment", Salt Lake City, UT, March1997.

Oremland, R. S. 1994. Biogeochemical transformations of selenium in anoxicenvironments. pp. 3889-419. In: W.T. Frankenberger, Jr. and S. Benson (eds.).Selenium in the Environment. Marcel Dekker, New York, NY.

Oswald, W. J., F. B. Green and T. L. Lundquist. 1998. Applications of advancedintegrated wastewater pond systems in California. Pres. The Asian Conf. Waterand Wastewater Management, Tehran, Iran, 2-4 March, 1998.

33

Packer, L. and A. Knight. 1986. Utilization of blue-green algal (cyanobacteria) toCleanse and Remove Se as an Option to Deal with Problems of Salinity andDrainage in the West Side of the San Joaquin Valley. Summarized in U.C.Salinity/Drainage Task Force Principal Accomplishments 1985-1990, Division ofAgriculture and Natural Resources, University of California, 1991, p.37.

Parker, D .R., C. Amrhein, and A. Thompson. 1997. Fate and partitioning of seleniumin a constructed wetland microcosm. pp.156-163 In UC Salinity/DrainageProgram Annual Report 1996-97, Division of Agriculture and Natural Resources,University of California.

Peterson, R. J. 1993. Composite reverse osmosis and nanofiltration membranes.J. Membrane Science 83:81-150.

Rael, R.M. and W.T. Frankenberger, Jr. 1995. Detection of selenomethionine in thefulvic fraction of a seleniferous sediment. Soil Biol. Biochem. 27:241-242.

Rael, R.M. and W.T. Frankenberger, Jr. 1996. Influence of pH, salinity, and seleniumon the growth of Aeromonas veronii in evaporation agricultural drainage water.Wat. Res. 30:422-430.

Rael, R. M.,E.C. Tuazon and W.T. Frankenberger, Jr. 1996. Gas-phase reactions ofdimethyl selenide with ozone and the hydroxyl and nitrate radicals. Atm. Environ.Environ. 30:1221-1232.

Reidel, G., J. Sanders, C. Gilmour, C. Goulden, R. Sanders, D. Breitburg, and R.Osman. 1997. Selenium interactions in aquatic ecosystems and effects on fishpopulations. Presentation at the Symposium "Understanding Selenium in theAquatic Environment", Salt Lake City, UT, March 1997.

Saiki, M. K. and T. P. Lowe. 1987. Selenium in aquatic organisms from subsurfaceagricultural drainage water, San Joaquin Valley, California. Arch. Environ.Contam. Toxicol. 16:16:657-670.

Schuler, C.A. 1987. Impacts of Agricultural Drainwater and Contamination onWetlands at Kesterson Reservoir, California. M.S. Thesis, Oregon StateUniversity, Corvallis, OR.

Schuler, C.A., R.G. Anthony and H.M. Ohlendorf. 1990. Selenium in wetlands andwaterfowl foods at Kesterson Reservoir, California, 1984. Arch. Environ. Contam.Toxicol. 19:845-853.

Skorupa, J. P., and H. M. Ohlendorf. 1991. Contaminants in drainage water and avianrisk thresholds. pp. 345-368. In A. Dinar and D. Zilberman (eds.) The Economyand Management of Water and Drainage in Agriculture. Kluwer AcademicPublishers, Norwell, MA.

34

Smith, B. E. 1992. Desalting and groundwater management in the San Joaquin Valley,California. Proc. of the NWSIA Biennial Conference, Newport Beach, CA.151-174.

Tanji, K. K. 1997. Application of TEAPS for the evaluation of selenium redox chemistryin the TLDD wetland system. pp.233-237. In UC Salinity/Drainage ProgramAnnual Report 1996-97, Division of Agriculture and Natural Resources,University of California.

Terry, N. 1997. Use of flow-through constructed wetlands for the remediation ofselenium in agricultural tiled-drainage water. UC Salinity/Drainage ProgramAnnual Report for 1996-1997. pp. 106-123. Centers for Water & WildlandResources.

Terry, N. and A. M. Zayed. 1994. Selenium volatilization by plants. pp. 343-367. In:W.T. Frankenberger, Jr. and S. Benson (eds.). Selenium in the Environment.Marcel Dekker, New York, NY.

Terry, N., C. Carlson, T. K. Raab and A. M. Zayed. 1992. Rates of seleniumvolatilization among crop species. J. Environ. Qual. 21: 341-344.

Thompson-Eagle, E.T. and W.T. Frankenberger, Jr. 1989a. Volatilization of seleniumfrom agricultural evaporation pond water. J. Environ. Qual. 19:125-131.

Thompson-Eagle, E.T., W.T. Frankenberger, Jr. and U. Karlson. 1989b. Volatilizationof selenium by Alternaria alternata. Appl. Environ. Microbiol. 55:1406-1413.

Thompson-Eagle, E. T. and W. T. Frankenberger, Jr. 1990. Protein-mediated seleniumbiomethylation in evaporation pond water. Environ. Toxicol.Chem. 9:1453-1462.

Thompson-Eagle, E.T. and W.T. Frankenberger, Jr. 1991a. Selenium biomethylation inan alkaline, saline environment. Wat. Res. 25:231-240.

Thompson-Eagle, E. T., W.T. Frankenberger, Jr. and K.E. Longley. 1991b. Removal ofselenium from agricultural drainage water through soil microbial transformations.pp. 169-186. In A. Dinar and D. Zilberman (eds.) The Economics andManagement of Water and Drainage in Agriculture. Kluwer Press.

Thompson-Eagle, E.T. and W. T. Frankenberger, Jr. 1992. bioremediation of soilscontaminated with selenium. pp. 261-310. In R. Lal and B.A. Stewart (eds.),Adv. Soil Sci. Springer-Verlag, New York, N.Y.

Zayed A. and N. Terry. 1994. Selenium volatilization in roots and shoots: effects ofshoot removal and sulfate level. J. Plant Physiol. 143: 8-14.

Zhang Y and J. Moore. 1997. Environmental conditions controlling seleniumvolatilization from a wetland system. Environ. Sci. Technol. 31: 511-517.

35

ACKNOWLEDGEMENTS

We would like to thank John Letey with the University of California Salinity andDrainage Program and Manucher Alemi of the Department of Water Resources,Sacramento, California, for their support of this project.

Tabl

e 1.

Man

agem

ent f

acto

rs a

ffect

ing

vola

tiliz

atio

n of

sel

eniu

m fr

om s

oil a

nd s

edim

ents

.

Man

agem

ent F

acto

rsR

efer

ence

sSe

leni

um v

olat

ilizat

ion

afte

r 37

days

was

incr

ease

d 1.

2-fo

ld a

t a C

/N =

20,

1.1

-fold

at C

/N =

10,

and

decr

ease

d by

11.

8 pe

rcen

t at C

/N =

5.

Nitr

ogen

app

licat

ions

of h

igh

rate

s ca

n in

hibi

t Se

vola

tiliz

atio

n.To

tal v

olat

ilizat

ion

over

21

days

was

incr

ease

d 2.

5-fo

ld w

ith C

o, 1

.67-

fold

with

Zn,

and

1.4

6-fo

ldw

ith N

i at 2

5 m

mol

kg-1

.

Karls

on a

ndFr

anke

nber

ger (

1988

)

Thre

e fu

ngal

isol

ates

incl

udin

g Ac

rem

oniu

m fa

lcifo

rme,

Pen

icilli

um c

itrin

um, a

nd U

locl

adiu

mtu

berc

ulat

um w

ere

isol

ated

from

sel

enife

rous

soi

ls a

ctiv

e in

Se

vola

tiliz

atio

n.W

ithou

t car

bon

amen

dmen

ts, v

olat

ilizat

ion

rate

s w

ere

up to

an

orde

r of m

agni

tude

hig

her w

ithSe

(IV) a

s th

e Se

sou

rce

whe

n co

mpa

red

to S

e(VI

).C

arbo

n ad

ditio

n in

the

form

of p

ectin

acc

eler

ated

Se

evol

utio

n 2-

to 1

30-fo

ld w

hich

was

mor

epr

onou

nced

with

Se(

VI).

With

car

bon

amen

dmen

ts, S

e(VI

) is

vola

tiliz

ed a

lmos

t as

rapi

dly

as S

e(IV

) fro

m a

ll so

ils.

With

pec

tin a

men

dmen

ts o

ver a

per

iod

of 1

18 d

ays,

tota

l Se

vola

tiliz

atio

n ra

nged

from

11.

3 to

51.

4pe

rcen

t of t

he a

dded

Se.

A m

inim

um S

e th

resh

old

for a

lkyl

sele

nide

pro

duct

ion

was

not

foun

d, a

pply

ing

Se a

dditi

ons

as lo

was

10

µg k

g-1 s

oil.

Karls

on a

ndFr

anke

nber

ger (

1989

)

A sa

line

(>20

dS

m-1

) sel

enife

rous

soi

l col

lect

ed fr

om P

ond

4, K

este

rson

Res

ervo

ir w

as a

ssay

ed fo

rD

MSe

pro

duct

ion

by in

cuba

ting

for u

p to

120

hou

rs a

t 22o C

. Opt

imum

con

ditio

ns fo

r Se

vola

tiliz

atio

n w

ere

obse

rved

at p

H 8

; moi

stur

e co

nten

t, fie

ld c

apac

ity; t

empe

ratu

re, 3

5o C; L

-m

ethi

onin

e, 1

00 m

g kg

-1 s

oil;

gala

ctur

onic

aci

d, 3

.6 g

C k

g-1 s

oils

; and

pro

tein

sou

rces

incl

udin

gca

sein

(2.0

g C

kg-1

soi

l) an

d al

bum

in (0

.05

to 2

.0 g

C k

g-1).

Albu

min

app

lied

to s

oil e

nhan

ced

gase

ous

Se p

rodu

ctio

n 19

.3-fo

ld o

ver t

he u

nam

ende

d co

ntro

l.

Fran

kenb

erge

r and

Karls

on (1

989)

37

The

phot

olys

is a

nd k

inet

ics

of D

MSe

with

OH

and

NO

3 rad

ical

s an

d O

3 was

det

erm

ined

. D

MSe

isno

t sub

ject

to p

hyto

lysi

s. R

ate

cons

tant

s fo

r rea

ctio

n w

ith O

H ra

dica

l and

O3 w

ere

6.78

x 1

0-11 a

nd6.

80 x

10-1

7 cm

3 mol

ecul

e-1 s-1

.Th

e ra

te c

onst

ant f

or th

e re

actio

n w

ith N

O3 r

adic

al w

as c

alcu

late

d at

1.4

x 1

0-11 c

m3 m

olec

ule-1

s-1.

Com

bini

ng th

ese

rate

con

stan

ts w

ith e

stim

ated

an

ambi

ent t

ropo

sphe

ric c

once

ntra

tions

of O

H a

ndN

O3 ra

dica

ls a

nd O

3, le

ads

to th

e fo

llow

ing

calc

ulat

ed li

fetim

es o

f DM

Se: O

H ra

dica

ls, 2

.7 h

; NO

3ra

dica

ls, 5

min

. and

O3,

5.8

h.

Atki

nson

et a

l. (1

990)

A se

dim

ent c

olum

n st

udy

esta

blis

hed

a Se

o (Se

vol

atiliz

atio

n po

tent

ial)

and

rate

con

stan

t (Se

vola

tiliz

atio

n ra

te c

oeffi

cien

t) de

rived

from

a fi

rst-o

rder

kin

etic

mod

el).

Incr

easi

ng th

e te

mpe

ratu

re fr

om 1

5 to

35o C

pro

mot

ed v

olat

ilizat

ion

of S

e (Q

10=1

.96)

The

amou

nt o

f Se

leac

hed

out o

f the

sed

imen

t was

inde

pend

ent o

f the

initi

al to

tal S

e in

vent

ory.

Vola

tiliz

atio

n of

Se

follo

wed

a fi

rst-o

rder

reac

tion

and

was

hig

hly

depe

nden

t on

the

initi

al S

eco

ncen

tratio

n.Th

e ad

ditio

n of

org

anic

am

endm

ents

pro

mot

ed v

olat

ilizat

ion

of th

e Se

bei

ng le

ss a

vaila

ble

for

leac

hing

. Th

e ap

plic

atio

n of

glu

ten,

a w

heat

sto

rage

pro

tein

, was

foun

d to

enh

ance

vol

atiliz

atio

nof

Se

1.7

to 3

.2-fo

ld o

ver t

he c

ontro

l mor

e st

rong

ly th

an a

ny o

ther

trea

tmen

t inc

ludi

ng c

asei

n,or

ange

pee

l, an

d ca

ttle

man

ure.

Cal

dero

ne e

t al.

(199

0)

With

Kes

ters

on s

edim

ent,

the

high

est S

e re

mov

al b

y vo

latil

izat

ion

resu

lted

from

the

com

bine

dap

plic

atio

n of

citr

us p

eel a

nd Z

nSO

4 (44

per

cent

) fol

low

ed b

y ci

trus

peel

alo

ne (4

0 pe

rcen

t), a

ndci

trus

peel

com

bine

d w

ith a

n Zn

SO4,

(NH

4)2 S

O4 a

nd A

crem

oniu

m fa

lcifo

rme

(30

perc

ent).

With

an

evap

orat

ion

pond

sed

imen

t obt

aine

d fro

m th

e Pe

ck R

anch

, the

hig

hest

am

ount

of S

ere

mov

ed w

as o

btai

ned

with

chi

tin (2

9 pe

rcen

t), c

attle

man

ure

(29

perc

ent)

and

citru

s pe

el a

lone

(27

perc

ent)

afte

r 273

day

s. W

ithou

t am

endm

ents

, 14.

0 pe

rcen

t of t

he n

ativ

e Se

was

vol

atiliz

ed in

273

days

.

Karls

on a

ndFr

anke

nber

ger (

1990

)

38

The

vapo

r pre

ssur

e of

vol

atile

Se

gase

s as

det

erm

ined

usi

ng a

n is

oten

isco

pe m

etho

d w

ere

calc

ulat

ed a

t 32.

0 an

d 0.

38 k

Pa fo

r DM

Se a

nd D

MD

Se, r

espe

ctiv

ely,

at 2

5o C.

The

enth

alpi

es o

f vap

oriz

atio

n w

ere

calc

ulat

ed a

t 31.

9 an

d 74

.9 k

J m

ol-1

, res

pect

ivel

y.Th

e so

lubi

lity

of D

MSe

is 0

.024

g/g

of w

ater

.Th

e H

enry

's L

aw c

onst

ant f

or D

MSe

was

cal

cula

ted

as 1

43 k

Pa k

g m

ol-1

(0.1

44 k

Pa m

3 mol

-1)

Karls

on e

t al.

(199

4)

Labo

rato

ry s

tudi

es re

veal

ed th

at s

elen

ium

vol

atiliz

atio

n ra

tes

wer

e hi

gher

und

er c

ontin

uous

ly m

oist

cond

ition

s (-3

3 kP

a) c

ompa

red

with

wet

ting

and

dryi

ng c

ycle

s.O

ptim

um a

ggre

gate

siz

e to

pro

mot

e vo

latil

izat

ion

of S

e w

as 0

.53

mm

whe

n co

mpa

red

to 0

.15,

1an

d 2

mm

.Th

e ap

plic

atio

n of

sal

ine

wel

l wat

er (7

.5 d

S m

-1 o

ver s

ix m

onth

s co

mpa

red

with

dei

oniz

ed w

ater

had

little

affe

ct o

n vo

latil

izat

ion

rate

s of

Se

from

a h

ighl

y sa

line

(22

dS m

-1) s

elen

ifero

us d

ewat

ered

sedi

men

t.

Fran

kenb

erge

r and

Karls

on (1

994b

)

A fie

ld s

tudy

dem

onst

ratin

g vo

latil

izat

ion

Se fr

om a

dew

ater

ed s

elen

ifero

us s

edim

ent r

evea

led

that

afte

r 22

mon

ths,

the

appl

icat

ion

of w

ater

plu

s til

lage

alo

ne re

mov

ed 3

2 pe

rcen

t of t

he S

e w

hile

cattl

e m

anur

e re

mov

ed 5

8 pe

rcen

t.Am

ong

the

para

met

ers

whi

ch e

nhan

ced

vola

tiliz

atio

n of

Se

wer

e an

ava

ilabl

e C

sou

rce,

aer

atio

n,m

oist

ure,

and

hig

h te

mpe

ratu

res.

Fran

kenb

erge

r and

Karls

on (1

995)

Sele

nom

ethi

onin

e w

as d

isco

vere

d in

the

fulv

ic fr

actio

n of

a s

elen

ifero

us s

edim

ent a

s de

tect

ed b

yse

lect

ed io

n-m

onito

ring

by G

C-M

S.R

ael a

ndFr

anke

nber

ger

(199

5)

39

Prod

ucts

of a

tmos

pher

ic im

porta

nt re

actio

ns o

f DM

Se re

veal

ed th

at d

imet

hyl s

elen

oxid

e[(C

H3)

2SeO

] is

a m

ajor

pro

duct

of t

he g

as-p

hase

reac

tion

of D

MSe

with

ozo

ne, w

ith a

n es

timat

edyi

eld

of a

ppro

xim

atel

y 90

per

cent

.Si

gnifi

cant

frag

men

tatio

n vi

a Se

-C b

ond

brea

kage

occ

urre

d du

ring

the

reac

tions

of D

MSe

with

the

OH

radi

cal a

nd th

e N

O3 ra

dica

l as

indi

cate

d by

HC

HO

form

atio

n yi

eldi

ng a

ppro

xim

atel

y 40

and

60

perc

ent,

resp

ectiv

ely.

Rae

l et a

l. (1

996)

A lo

ng-te

rm fi

eld

expe

rimen

t was

car

ried

out a

t Kes

ters

on R

eser

voir

to d

eter

min

e op

timum

man

agem

ent f

acto

rs re

mov

ing

sele

nium

from

soi

l.O

ver a

per

iod

of 1

00 m

onth

s, 6

8 to

88

perc

ent o

f the

tota

l sel

eniu

m w

as d

issi

pate

d fro

m th

e to

psoi

l(0

-15

cm).

The

patte

rn o

f Se

depl

etio

n in

soi

l was

not

cor

rela

ted

with

rain

fall

even

ts n

or w

ith te

mpe

ratu

re.

The

high

est a

mou

nt o

f Se

depl

etio

n oc

curre

d w

ith th

e am

endm

ent o

f the

pro

tein

, cas

ein.

A tw

o-co

mpa

rtmen

t mod

el w

as s

uper

ior t

o a

one

com

partm

ent m

odel

for d

escr

ibin

g th

e lo

ng-te

rmki

netic

s of

Se

depl

etio

n in

the

soil.

The

rate

of S

e de

plet

ion

was

initi

ally

fast

er th

an a

t lat

er ti

mes

indi

catin

g th

at a

rate

lim

iting

mec

hani

sm c

hang

ed d

urin

g th

e tim

e of

the

stud

y.

Flur

y et

al.

(199

7)

40

Tabl

e 2.

Man

agem

ent f

acto

rs a

ffect

ing

vola

tiliz

atio

n of

sel

eniu

m fr

om w

ater

.

Man

agem

ent F

acto

rsR

efer

ence

sN

atur

al fo

rmat

ion

of D

MSe

in e

vapo

ratio

n po

nd w

ater

was

less

than

1 p

erce

nt o

f the

tota

l Se

inve

ntor

y af

ter 4

0 da

ys o

f inc

ubat

ion.

No

Se m

ethy

latio

n to

ok p

lace

in a

utoc

lave

, una

men

ded

pond

wat

er.

L-M

ethi

onin

e (1

0 µM

) stim

ulat

ed D

MSe

pro

duct

ion

in n

onst

erile

eva

pora

tion

pond

wat

er.

Incr

easi

ng th

e te

mpe

ratu

re to

35o C

and

the

addi

tion

of 1

per

cent

glu

cose

with

a fu

ngal

inoc

ulum

,Al

tern

aria

alte

rnat

a, d

oubl

ed D

MSe

pro

duct

ion

over

the

cont

rols

afte

r 25

days

of i

ncub

atio

n.C

arbo

n so

urce

s su

ch a

s gl

ucos

e, m

alto

se, s

ucro

se, a

nd g

alac

turo

nic

acid

at 2

g C

L-1

und

eram

bien

t con

ditio

ns s

light

ly e

nhan

ced

indi

geno

us S

e vo

latil

izat

ion

(1.5

-fold

).O

f the

am

ino

acid

s te

sted

, L-m

ethi

onin

e (0

.02

g C

L-1

) stim

ulat

ed th

e D

MSe

evo

lutio

n fro

m p

ond

wat

er m

ore

so th

an L

-cys

tein

e, L

- cys

tine,

and

L-s

erin

e.Th

e pr

otei

ns, e

gg a

lbum

in, c

asei

n, a

nd g

lute

n (2

g C

L-1

) dra

mat

ical

ly in

crea

sed

Se v

olat

ilizat

ion

caus

ing

a 23

, 41,

and

10

perc

ent S

e lo

ss fr

om th

e in

vent

ory,

resp

ectiv

ely

afte

r 43

days

of

incu

batio

n.

Thom

pson

-Eag

le a

ndFr

anke

nber

ger (

1989

a).

Alte

rnar

ia a

ltern

ata

was

isol

ated

from

wat

er s

ampl

es c

olle

cted

from

eva

pora

tion

pond

faci

litie

s in

the

San

Joaq

uin

Valle

y.Th

e op

timum

pH

and

tem

pera

ture

for m

ethy

latio

n of

Se

wer

e 6.

5 an

d 30

o C, r

espe

ctiv

ely.

Sele

nate

and

sel

enite

wer

e m

ethy

late

d m

ore

rapi

dly

than

sel

eniu

m s

ulfid

e an

d va

rious

org

anic

Se

com

poun

ds (6

-sel

enog

uano

sine

, 6-s

elen

oino

sine

, sel

eno-

DL-

met

hion

ine,

6-s

elen

o pu

rine)

.

L-M

ethi

onin

e an

d m

ethy

l cob

alam

ine

(0.1

µM

) stim

ulat

ed D

MSe

pro

duct

ion.

Thom

pson

-Eag

le e

t al.

(198

9b)

41

Labo

rato

ry s

tudi

es s

how

ed th

at S

e vo

latil

izat

ion

is p

rote

in-p

eptid

e lim

ited

rath

er th

an n

itrog

en-,

amin

o ac

id-,

or c

arbo

n-lim

ited.

Prot

ein

sour

ces

incl

udin

g co

ttons

eed

mea

l, ch

eese

whe

y an

d ye

ast s

ludg

e dr

amat

ical

ly in

crea

sed

Se v

olat

ilizat

ion

(29-

fold

, 300

-fold

and

41-

fold

, res

pect

ivel

y) o

ver u

nam

ende

d w

ater

sam

ples

, with

over

10

perc

ent o

f the

Se

inve

ntor

y co

nver

ted

into

DM

Se in

17

days

.A

sing

le c

asei

n am

endm

ent (

0.2

g C

L-1

) to

wat

er c

olum

ns in

the

field

cau

sed

a 38

per

cent

Se

loss

from

the

initi

al in

vent

ory

of a

San

Joa

quin

Val

ley

evap

orat

ion

pond

in 1

42 d

ays.

Thom

pson

-Eag

le a

ndFr

anke

nber

ger (

1990

a)

Biom

ethy

latio

n w

as o

ptim

al in

wel

l-mix

ed, a

erob

ic s

yste

m a

men

ded

with

a p

rote

in s

ourc

e.

Cof

acto

rs (1

0 µM

hom

ocys

tein

e an

d 10

µM

redu

ced

glut

athi

one)

enh

ance

d pr

oduc

tion

of D

MSe

inpe

pton

e-am

ende

d po

nd w

ater

.

The

spec

ies

of in

orga

nic

Se p

rese

nt,

−2 3SeO

and

−2 4

SeO

had

littl

e af

fect

on

the

met

hyla

tion

effic

ienc

y.In

crea

sing

the

Se c

once

ntra

tion

to 2

, 4, 1

2, 2

2, o

r 102

mg

Se L

-1 in

pep

tone

-am

ende

d po

nd w

ater

decr

ease

d th

e pe

rcen

tage

of S

e re

mov

ed fr

om 8

to 4

, 2, 1

.3, 0

.7 a

nd 0

.1 p

erce

nt, r

espe

ctiv

ely.

The

addi

tion

of c

asei

n (4

g L

-1) i

ncre

ased

bac

teria

l num

bers

1,0

00-fo

ld a

nd s

timul

ated

Se

vola

tiliz

atio

n 25

-fold

.

Thom

pson

-Eag

le a

ndFr

anke

nber

ger (

1991

a).

Aero

mon

as v

eron

ii, is

olat

ed fr

om a

sel

enife

rous

agr

icul

tura

l dra

inag

e w

ater

, met

abol

ized

pep

tone

and

was

act

ive

in v

olat

ilizin

g Se

prin

cipa

lly a

s D

MSe

.O

ther

vol

atile

pro

duct

s in

clud

ed d

imet

hyl d

isul

fide,

met

hyl s

elen

ol, d

imet

hyl s

elen

osul

fide,

and

dim

ethy

l dis

elen

ide.

The

rate

of S

e vo

latil

izat

ion

corre

late

d w

ith th

e gr

owth

of A

. ver

onii

with

the

high

est l

evel

s of

DM

Sere

leas

ed d

urin

g th

e ex

pone

ntia

l pha

se o

f gro

wth

.

Rae

l and

Fra

nken

berg

er(1

996)

G:\a

lem

i\tas

k2a.

doc